Systems and methods for decoding image files containing depth maps stored as metadata

[0001] The present invention is a continuation of U.S. patent application Ser. No. 15/396,024 filed Dec. 30, 2016, which is a continuation of U.S. patent application Ser. No. 14/667,492 filed Mar. 24, 2015 and issued as U.S. Pat. No. 9,536,166 on Jan. 3, 2017, which is a continuation of U.S. patent application Ser. No. 14/504,687, filed Oct. 2, 2014 and issued as U.S. Pat. No. 9,042,667 on May 26, 2015, which is a continuation of U.S. patent application Ser. No. 14/477,374, filed Sept. 4, 2014, and issued as U.S. Pat. No. 9,036,931 on May 19, 2015, which is a continuation of U.S. patent application Ser. No. 13/955,411, filed Jul. 31, 2013 and issued as U.S. Pat. No. 8,831,367 on Sep. 9, 2014, which is a continuation of U.S. patent application Ser. No. 13/631,736, filed Sep. 28, 2012 and issued as U.S. Pat. No. 8,542,933 on Sep. 24, 2013, which claims priority to U.S. Provisional Application No. 61/540,188 entitled “JPEG-DX: A Backwards-compatible, Dynamic Focus Extension to JPEG”, to Venkataraman et al., filed Sep. 28, 2011, the disclosures of which are incorporated herein by reference in their entirety.

[0002] The present invention relates to decoding of image files and more specifically to the decoding of light field image files.

[0003] The ISO/IEC 10918-1 standard, more commonly referred to as the JPEG standard after the Joint Photographic Experts Group that developed the standard, establishes a standard process for digital compression and coding of still images. The JPEG standard specifies a codec for compressing an image into a bitstream and for decompressing the bitstream back into an image.

[0004] A variety of container file formats including the JPEG File Interchange Format (JFIF) specified in ISO/IEC 10918-5 and the Exchangeable Image File Format (Exif) and can be used to store a JPEG bitstream. JFIF can be considered a minimal file format that enables JPEG bitstreams to be exchanged between a wide variety of platforms and applications. The color space used in JFIF files is YCbCr as defined by CCIR Recommendation 601, involving 256 levels. The Y, Cb, and Cr components of the image file are converted from R, G, and B, but are normalized so as to occupy the full 256 levels of an 8-bit binary encoding. YCbCr is one of the compression formats used by JPEG. Another popular option is to perform compression directly on the R, G and B color planes. Direct RGB color plane compression is also popular when lossless compression is being applied.

[0005] A JPEG bitstream stores 16-bit word values in big-endian format. JPEG data in general is stored as a stream of blocks, and each block is identified by a marker value. The first two bytes of every JPEG bitstream are the Start Of Image (SOI) marker values FFh D8h. In a JFIF-compliant file there is a JFIF APP0 (Application) marker, immediately following the SOI, which consists of the marker code values FFh E0h and the characters JFIF in the marker data, as described in the next section. In addition to the JFIF marker segment, there may be one or more optional JFIF extension marker segments, followed by the actual image data.

[0006] Overall, the JFIF format supports sixteen “Application markers” to store metadata. Using application markers makes it is possible for a decoder to parse a JFIF file and decode only required segments of image data. Application markers are limited to 64K bytes each but it is possible to use the same maker ID multiple times and refer to different memory segments.

[0007] An APP0 marker after the SOI marker is used to identify a JFIF file. Additional APP0 marker segments can optionally be used to specify JFIF extensions. When a decoder does not support decoding a specific JFIF application marker, the decoder can skip the segment and continue decoding.

[0008] One of the most popular file formats used by digital cameras is Exif. When Exif is employed with JPEG bitstreams, an APP1 Application marker is used to store the Exif data. The Exif tag structure is borrowed from the Tagged Image File Format (TIFF) maintained by Adobe Systems Incorporated of San Jose, Calif.

[0009] Systems and methods in accordance with embodiments of the invention are configured to render images using light field image files containing an image synthesized from light field image data and metadata describing the image that includes a depth map. One embodiment of the invention includes a processor and memory containing a rendering application and a light field image file including an encoded image and metadata describing the encoded image, where the metadata comprises a depth map that specifies depths from the reference viewpoint for pixels in the encoded image. In addition, the rendering application configures the processor to: locate the encoded image within the light field image file; decode the encoded image; locate the metadata within the light field image file; and post process the decoded image by modifying the pixels based on the depths indicated within the depth map to create a rendered image.

[0010] In a further embodiment the rendering application configuring the processor to post process the decoded image by modifying the pixels based on the depths indicated within the depth map to create the rendered image comprises applying a depth based effect to the pixels of the decoded image.

[0011] In another embodiment, the depth based effect comprises at least one effect selected from the group consisting of: modifying the focal plane of the decoded image; modifying the depth of field of the decoded image; modifying the blur in out-of-focus regions of the decoded image; locally varying the depth of field of the decoded image; creating multiple focus areas at different depths within the decoded image; and applying a depth related blur.

[0012] In a still further embodiment, the encoded image is an image of a scene synthesized from a reference viewpoint using a plurality of lower resolution images that capture the scene from different viewpoints, the metadata in the light field image file further comprises pixels from the lower resolution images that are occluded in the reference viewpoint, and the rendering application configuring the processor to post process the decoded image by modifying the pixels based on the depths indicated within the depth map to create the rendered image comprises rendering an image from a different viewpoint using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint.

[0013] In still another embodiment, the metadata in the light field image file includes descriptions of the pixels from the lower resolution images that are occluded in the reference viewpoint including the color, location, and depth of the occluded pixels, and rendering an image from a different viewpoint using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint further includes: shifting pixels from the decoded image and the occluded pixels in the metadata to the different viewpoint based upon the depths of the pixels; determining pixel occlusions; and generating an image from the different viewpoint using the shifted pixels that are not occluded and by interpolating to fill in missing pixels using adjacent pixels that are not occluded.

[0014] In a yet further embodiment, the image rendered from the different viewpoint is part of a stereo pair of images.

[0015] In yet another embodiment, the metadata in the light field image file further comprises a confidence map for the depth map, where the confidence map indicates the reliability of the depth values provided for pixels by the depth map, and rendering an image from a different viewpoint using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint further comprises applying at least one filter based upon the confidence map.

[0016] In a further embodiment again, the metadata in the light field image file further comprises an edge map that indicates pixels in the decoded image that lie on a discontinuity, and rendering an image from a different viewpoint using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint further comprises applying at least one filter based upon the edge map.

[0017] In another embodiment again, the edge map identifies whether a pixel lies on an intensity discontinuity.

[0018] In a further additional embodiment, the edge map identifies whether a pixel lies on an intensity and depth discontinuity.

[0019] In another additional embodiment, the metadata in the light field image file further comprises a missing pixel map that indicates pixels in the decoded image that do not correspond to a pixel from the plurality of low resolution images of the scene and that are generated by interpolating pixel values from adjacent pixels in the synthesized image, and rendering an image from a different viewpoint using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint further comprises ignoring pixels based upon the missing pixel map.

[0020] In a still further embodiment again, the light field image file conforms to the JFIF standard and the encoded image is encoded in accordance with the JPEG standard, the memory comprises a JPEG decoder application, and the rendering application configures the processor to: locate the encoded image by locating a Start of Image marker within the light field image file; and decode the encoded image using the JPEG decoder.

[0021] In still another embodiment again, the metadata is located within an Application marker segment within the light field image file.

[0022] In a still further additional embodiment, the Application marker segment is identified using the APP9 marker.

[0023] In still another additional embodiment, the depth map is encoded in accordance with the JPEG standard using lossless compression, and the rendering application configures the processor to: locate at least one Application marker segment containing the metadata comprising the depth map; and decode the depth map using the JPEG decoder.

[0024] In a yet further embodiment again, the encoded image is an image of a scene synthesized from a reference viewpoint using a plurality of lower resolution images that capture the scene from different viewpoints, the metadata in the light field image file further comprises pixels from the lower resolution images that are occluded in the reference viewpoint, the rendering application configures the processor to locate at least one Application marker segment containing the metadata comprising the pixels from the lower resolution images that are occluded in the reference viewpoint, and the rendering application configuring the processor to post process the decoded image by modifying the pixels based on the depth of the pixel indicated within the depth map to create the rendered image comprises rendering an image from a different viewpoint using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint.

[0025] In yet another embodiment again, the metadata in the light field image file includes descriptions of the pixels from the lower resolution images that are occluded in the reference viewpoint including the color, location, and depth of the occluded pixels, and rendering an image from a different viewpoint using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint further includes: shifting pixels from the decoded image and the occluded pixels in the metadata to the different viewpoint based upon the depths of the pixels; determining pixel occlusions; and generating an image from the different viewpoint using the shifted pixels that are not occluded and by interpolating to fill in missing pixels using adjacent pixels that are not occluded.

[0026] In a yet further additional embodiment, the image rendered from the different viewpoint is part of a stereo pair of images.

[0027] In yet another additional embodiment, the metadata in the light field image file further comprises a confidence map for the depth map, where the confidence map indicates the reliability of the depth values provided for pixels by the depth map, and rendering an image from a different viewpoint using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint further comprises applying at least one filter based upon the confidence map.

[0028] In a further additional embodiment again, the metadata in the light field image file further comprises an edge map that indicates pixels in the decoded image that lie on a discontinuity, and rendering an image from a different viewpoint using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint further comprises applying at least one filter based upon the edge map.

[0029] In another additional embodiment again, the edge map identifies whether a pixel lies on an intensity discontinuity.

[0030] In a still yet further embodiment again, the edge map identifies whether a pixel lies on an intensity and depth discontinuity.

[0031] In still yet another embodiment again, the edge map is encoded in accordance with the JPEG standard using lossless compression, and the rendering application configures the processor to: locate at least one Application marker segment containing the metadata comprising the edge map; and decode the edge map using the JPEG decoder.

[0032] In a still yet further additional embodiment, the metadata in the light field image file further comprises a missing pixel map that indicates pixels in the decoded image that do not correspond to a pixel from the plurality of low resolution images of the scene and that are generated by interpolating pixel values from adjacent pixels in the synthesized image, and rendering an image from a different viewpoint using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint further comprises ignoring pixels based upon the missing pixel map.

[0033] In still yet another additional embodiment, the missing pixel map is encoded in accordance with the JPEG standard using lossless compression, and the rendering application configures the processor to: locate at least one Application marker segment containing the metadata comprising the missing pixel; and decode the missing pixel map using the JPEG decoder.

[0034] An embodiment of the method of the invention includes locating an encoded image within an light field image file using a rendering device, decoding the encoded image using the rendering device, locating the metadata within the light field image file using the rendering device, and post processing the decoded image by modifying the pixels based on the depths indicated within the depth map to create a rendered image using the rendering device.

[0035] In a further embodiment of the method of the invention, post processing the decoded image by modifying the pixels based on the depths indicated within the depth map to create the rendered image comprises applying a depth based effect to the pixels of the decoded image using the rending device.

[0036] In another embodiment of the method of the invention, the depth based effect comprises at least one effect selected from the group consisting of: modifying the focal plane of the decoded image using the rendering device; modifying the depth of field of the decoded image using the rendering device; modifying the blur in out-of-focus regions of the decoded image using the rendering device; locally varying the depth of field of the decoded image using the rendering device; creating multiple focus areas at different depths within the decoded image using the rendering device; and applying a depth related blur using the rendering device.

[0037] In a still further embodiment of the method of the invention, the encoded image is an image of a scene synthesized from a reference viewpoint using a plurality of lower resolution images that capture the scene from different viewpoints, the metadata in the light field image file further comprises pixels from the lower resolution images that are occluded in the reference viewpoint, and post processing the decoded image by modifying the pixels based on the depths indicated within the depth map to create the rendered image comprises using the depth map and the pixels from the lower resolution images that are occluded in the reference viewpoint to render an image from a different viewpoint using the rendering device.

[0038] Another further embodiment of the invention includes a machine readable medium containing processor instructions, where execution of the instructions by a processor causes the processor to perform a process involving: locating an encoded image within a light field image file, where the light field image file includes an encoded image and metadata describing the encoded image comprising a depth map that specifies depths from the reference viewpoint for pixels in the encoded image; decoding the encoded image; locating the metadata within the light field image file; and post processing the decoded image by modifying the pixels based on the depths indicated within the depth map to create a rendered image.

[0063] Turning now to the drawings, systems and methods for storing images synthesized from light field image data and metadata describing the images in electronic files and for rendering images using the stored image and the metadata in accordance with embodiments of the invention are illustrated. A file containing an image synthesized from light field image data and metadata derived from the light field image data can be referred to as a light field image file. As is discussed further below, the encoded image in a light field image file is typically synthesized using a super resolution process from a number of lower resolution images. The light field image file can also include metadata describing the synthesized image derived from the light field image data that enables post processing of the synthesized image. In many embodiments, a light field image file is created by encoding an image synthesized from light field image data and combining the encoded image with a depth map derived from the light field image data. In several embodiments, the encoded image is synthesized from a reference viewpoint and the metadata includes information concerning pixels in the light field image that are occluded from the reference viewpoint. In a number of embodiments, the metadata can also include additional information including (but not limited to) auxiliary maps such as confidence maps, edge maps, and missing pixel maps that can be utilized during post processing of the encoded image to improve the quality of an image rendered using the light field image data file.

[0064] In many embodiments, the light field image file is compatible with the JPEG File Interchange Format (JFIF). The synthesized image is encoded as a JPEG bitstream and stored within the file. The accompanying depth map, occluded pixels and/or any appropriate additional information including (but not limited to) auxiliary maps are then stored within the JFIF file as metadata using an Application marker to identify the metadata. A legacy rendering device can simply display the synthesized image by decoding the JPEG bitstream. Rendering devices in accordance with embodiments of the invention can perform additional post-processing on the decoded JPEG bitstream using the depth map and/or any available auxiliary maps. In many embodiments, the maps included in the metadata can also be compressed using lossless JPEG encoding and decoded using a JPEG decoder. Although much of the discussion that follows references the JFIF and JPEG standards, these standards are simply discussed as examples and it should be appreciated that similar techniques can be utilized to embed metadata derived from light field image data used to synthesize the encoded image within a variety of standard file formats, where the synthesized image and/or maps are encoded using any of a variety of standards based image encoding processes.

[0065] By transmitting a light field image file including an encoded image, and metadata describing the encoded image, a rendering device (i.e. a device configured to generate an image rendered using the information within the light field image file) can render new images using the information within the file without the need to perform super resolution processing on the original light field image data. In this way, the amount of data transmitted to the rendering device and the computational complexity of rendering an image is reduced. In several embodiments, rendering devices are configured to perform processes including (but not limited to) refocusing the encoded image based upon a focal plane specified by the user, synthesizing an image from a different viewpoint, and generating a stereo pair of images. The capturing of light field image data and the encoding and decoding of light field image files in accordance with embodiments of the invention are discussed further below.

Capturing Light Field Image Data

[0066] A light field, which is often defined as a 4D function characterizing the light from all direction at all points in a scene, can be interpreted as a two-dimensional (2D) collection of 2D images of a scene. Array cameras, such as those described in U.S. patent application Ser. No. 12/935,504 entitled “Capturing and Processing of Images using Monolithic Camera Array with Heterogeneous Imagers” to Venkataraman et al., can be utilized to capture light field images. In a number of embodiments, super resolution processes such as those described in U.S. patent application Ser. No. 12/967,807 entitled “Systems and Methods for Synthesizing High Resolution Images Using Super-Resolution Processes” to Lelescu et al., are utilized to synthesize a higher resolution 2D image or a stereo pair of higher resolution 2D images from the lower resolution images in the light field captured by an array camera. The terms high or higher resolution and low or lower resolution are used here in a relative sense and not to indicate the specific resolutions of the images captured by the array camera. The disclosures of U.S. patent application Ser. 12/935,504 and U.S. patent application Ser. No. 12/967,807 are hereby incorporated by reference in their entirety.

[0067] Each two-dimensional (2D) image in a captured light field is from the viewpoint of one of the cameras in the array camera. A high resolution image synthesized using super resolution processing is synthesized from a specific viewpoint that can be referred to as a reference viewpoint. The reference viewpoint can be from the viewpoint of one of the cameras in a camera array. Alternatively, the reference viewpoint can be an arbitrary virtual viewpoint.

[0068] Due to the different viewpoint of each of the cameras, parallax results in variations in the position of foreground objects within the images of the scene. Processes for performing parallax detection are discussed in U.S. Provisional Patent Application Ser. No. 61/691,666 entitled “Systems and Methods for Parallax Detection and Correction in Images Captured Using Array Cameras” to Venkataraman et al., the disclosure of which is incorporated by reference herein in its entirety. As is disclosed in U.S. Provisional Patent Application Ser. No. 61/691,666, a depth map from a reference viewpoint can be generated by determining the disparity between the pixels in the images within a light field due to parallax. A depth map indicates the distance of the surfaces of scene objects from a reference viewpoint. In a number of embodiments, the computational complexity of generating depth maps is reduced by generating an initial low resolution depth map and then increasing the resolution of the depth map in regions where additional depth information is desirable such as (but not limited to) regions involving depth transitions and/or regions containing pixels that are occluded in one or more images within the light field.

[0069] During super resolution processing, a depth map can be utilized in a variety of ways. U.S. patent application Ser. No. 12/967,807 describes how a depth map can be utilized during super resolution processing to dynamically refocus a synthesized image to blur the synthesized image to make portions of the scene that do not lie on the focal plane to appear out of focus. U.S. patent application Ser. No. 12/967,807 also describes how a depth map can be utilized during super resolution processing to generate a stereo pair of higher resolution images for use in 3D applications. A depth map can also be utilized to synthesize a high resolution image from one or more virtual viewpoints. In this way, a rendering device can simulate motion parallax and a dolly zoom (i.e. virtual viewpoints in front or behind the reference viewpoint). In addition to utilizing a depth map during super-resolution processing, a depth map can be utilized in a variety of post processing processes to achieve effects including (but not limited to) dynamic refocusing, generation of stereo pairs, and generation of virtual viewpoints without performing super-resolution processing. Light field image data captured by array cameras, storage of the light field image data in a light field image file, and the rendering of images using the light field image file in accordance with embodiments of the invention are discussed further below.

Array Camera Architecture

[0070] Array cameras in accordance with embodiments of the invention are configured so that the array camera software can control the capture of light field image data and can capture the light field image data into a file that can be used to render one or more images on any of a variety of appropriately configured rendering devices. An array camera including an imager array in accordance with an embodiment of the invention is illustrated in FIG. 1. The array camera 100 includes a sensor 102 having an array of focal planes 104 and which is configured to communicate with a processor 108. The processor is also configured to communicate with one or more different types of memory 110 that can be utilized to store image data and/or contain machine readable instructions utilized to configure the processor to perform processes including (but not limited to) the various processes described below. The array camera 100 also includes a display 112 that can be utilized by the processor 108 to present a user interface to a user and to display an image rendered using the light field image data. Although the processor is illustrated as a single processor, array cameras in accordance with embodiments of the invention can utilize a single processor or multiple processors including (but not limited to) a graphics processing unit (GPU).

[0071] In the illustrated embodiment, the processor receives image data generated by the sensor and reconstructs the light field captured by the sensor from the image data. The processor can manipulate the light field in any of a variety of different ways including (but not limited to) determining the depth and visibility of the pixels in the light field and synthesizing higher resolution 2D images from the image data of the light field. Sensors including multiple focal planes are discussed in U.S. patent application Ser. No. 13/106,797 entitled “Architectures for System on Chip Array Cameras”, to Pain et al., the disclosure of which is incorporated herein by reference in its entirety.

[0072] In the illustrated embodiment, the focal planes are configured in a 5×5 array. Each focal plane 104 on the sensor is capable of capturing an image of the scene. The sensor elements utilized in the focal planes can be individual light sensing elements such as, but not limited to, traditional CIS (CMOS Image Sensor) pixels, CCD (charge-coupled device) pixels, high dynamic range sensor elements, multispectral sensor elements and/or any other structure configured to generate an electrical signal indicative of light incident on the structure. In many embodiments, the sensor elements of each focal plane have similar physical properties and receive light via the same optical channel and color filter (where present). In other embodiments, the sensor elements have different characteristics and, in many instances, the characteristics of the sensor elements are related to the color filter applied to each sensor element.

[0073] In many embodiments, an array of images (i.e. a light field) is created using the image data captured by the focal planes in the sensor. As noted above, processors 108 in accordance with many embodiments of the invention are configured using appropriate software to take the image data within the light field and synthesize one or more high resolution images. In several embodiments, the high resolution image is synthesized from a reference viewpoint, typically that of a reference focal plane 104 within the sensor 102. In many embodiments, the processor is able to synthesize an image from a virtual viewpoint, which does not correspond to the viewpoints of any of the focal planes 104 in the sensor 102. Unless all of the objects within a captured scene are a significant distance from the array camera, the images in the light field will include disparity due to the different fields of view of the focal planes used to capture the images. Processes for detecting and correcting for disparity when performing super-resolution processing in accordance with embodiments of the invention are discussed in U.S. Provisional Patent Application Ser. No. 61/691,666 (incorporated by reference above). The detected disparity can be utilized to generate a depth map. The high resolution image and depth map can be encoded and stored in memory 110 in a light field image file. The processor 108 can use the light field image file to render one or more high resolution images. The processor 108 can also coordinate the sharing of the light field image file with other devices (e.g. via a network connection), which can use the light field image file to render one or more high resolution images.

[0074] Although a specific array camera architecture is illustrated in FIG. 1, alternative architectures can also be utilized in accordance with embodiments of the invention. Systems and methods for encoding high resolution images and depth maps for storage in electronic files in accordance with embodiments of the invention are discussed below.

Capturing and Storing Light Field Image Data

[0075] Processes for capturing and storing light field image data in accordance with many embodiments of the invention involve capturing light field image data, generating a depth map from a reference viewpoint, and using the light field image data and the depth map to synthesize an image from the reference viewpoint. The synthesized image can then be compressed for storage. The depth map and additional data that can be utilized in the post processing can also be encoded as metadata that can be stored in the same container file with the encoded image.

[0076] A process for capturing and storing light field image data in accordance with an embodiment of the invention is illustrated in FIG. 2. The process 200 includes capturing (202) light field image data. In several embodiments, the light field image data is captured using an array camera similar to the array cameras described above. In other embodiments, any of a variety of image capture device(s) can be utilized to capture light field image data. The light field image data is used to generate (204) a depth map. A depth map can be generated using any of a variety of techniques including (but not limited to) using any of the processes disclosed in U.S. Provisional Patent Application Ser. No. 61/691,666 or U.S. patent application Ser. No. 13/623,091 entitled “Systems and Methods for Determining Depth from Multiple Views of a Scene that Include Aliasing Using Hypothesized Fusion”, to Venkatarman et al. The disclosure of U.S. patent Ser. No. 13/623,091 is incorporated by reference herein in its entirety.

[0077] The light field image data and the depth map can be utilized to synthesize (206) an image from a specific viewpoint. In many embodiments, the light field image data includes a number of low resolution images that are used to synthesize a higher resolution image using a super resolution process. In a number of embodiments, a super resolution process such as (but not limited to) any of the super resolution processes disclosed in U.S. patent application Ser. No. 12/967,807 can be utilized to synthesize a higher resolution image from the reference viewpoint.

[0078] In order to be able to perform post processing to modify the synthesized image without the original light field image data, metadata can be generated (208) from the light field image data, the synthesized image, and/or the depth map. The metadata data can be included in a light field image file and utilized during post processing of the synthesized image to perform processing including (but not limited to) refocusing the encoded image based upon a focal plane specified by the user, and synthesizing one or more images from a different viewpoint. In a number of embodiments, the auxiliary data includes (but is not limited to) pixels in the light field image data occluded from the reference viewpoint used to synthesize the image from the light field image data, one or more auxiliary maps including (but not limited to) a confidence map, an edge map, and/or a missing pixel map. Auxiliary data that is formatted as maps or layers provide information corresponding to pixel locations within the synthesized image. A confidence map is produced during the generation of a depth map and reflects the reliability of the depth value for a particular pixel. This information may be used to apply different filters in areas of the image and improve image quality of the rendered image. An edge map defines which pixels are edge pixels, which enables application of filters that refine edges (e.g. post sharpening). A missing pixel map represents pixels computed by interpolation of neighboring pixels and enables selection of post-processing filters to improve image quality. As can be readily appreciated, the specific metadata generated depends upon the post processing supported by the image data file. In a number of embodiments, no auxiliary data is included in the image data file.

[0079] In order to generate an image data file, the synthesized image is encoded (210). The encoding typically involves compressing the synthesized image and can involve lossless or lossy compression of the synthesized image. In many embodiments, the depth map and any auxiliary data are written (212) to a file with the encoded image as metadata to generate a light field image data file. In a number of embodiments, the depth map and/or the auxiliary maps are encoded. In many embodiments, the encoding involves lossless compression.

[0080] Although specific processes for encoding light field image data for storage in a light field image file are discussed above, any of a variety of techniques can be utilized to process light field image data and store the results in an image file including but not limited to processes that encode low resolution images captured by an array camera and calibration information concerning the array camera that can be utilized in super resolution processing. Storage of light field image data in JFIF files in accordance with embodiments of the invention is discussed further below.

Image Data Formats

[0081] In several embodiments, the encoding of a synthesized image and the container file format utilized to create the light field image file are based upon standards including but not limited to the JPEG standard (ISO/IEC 10918-1) for encoding a still image as a bitstream and the JFIF standard (ISO/IEC 10918-5). By utilizing these standards, the synthesized image can be rendered by any rendering device configured to support rendering of JPEG images contained within JFIF files. In many embodiments, additional data concerning the synthesized image such as (but not limited to) a depth map and auxiliary data that can be utilized in the post processing of the synthesized image can be stored as metadata associated with an Application marker within the JFIF file. Conventional rendering devices can simply skip Application markers containing this metadata. Rendering device in accordance with many embodiments of the invention can decode the metadata and utilize the metadata in any of a variety of post processing processes.

[0082] A process for encoding an image synthesized using light field image data in accordance with the JPEG specification and for including the encoded image and metadata that can be utilized in the post processing of the image in a JFIF file in accordance with an embodiment of the invention is illustrated in FIG. 3. The process 300 includes encoding (302) an image synthesized from light field image data in accordance with the JPEG standard. The image data is written (304) to a JFIF file. A depth map for the synthesized image is compressed (306) and the compressed depth map and any additional auxiliary data are written (308) as metadata to an Application marker segment of the JFIF file containing the encoded image. Where the auxiliary data includes maps, the maps can also be compressed by encoding the maps in accordance with the JPEG standard. At which point, the JFIF file contains an encoded image and metadata that can be utilized to perform post processing on the encoded image in ways that utilize the additional information captured in the light field image data utilized to synthesize the high resolution image (without the need to perform super resolution processing on the underlying light field image data).

[0083] Although specific processes are discussed above for storing light field image data in JFIF files, any of a variety of processes can be utilized to encode synthesized images and additional metadata derived from the light field image data used to synthesize the encoded images in a JFIF file as appropriate to the requirements of a specific application in accordance with embodiments of the invention. The encoding of synthesized images and metadata for insertion into JFIF files in accordance with embodiments of the invention are discussed further below. Although much of the discussion that follows relates to JFIF files, synthesized images and metadata can be encoded for inclusion in a light field image file using any of a variety of proprietary or standards based encoding techniques and/or utilizing any of a variety of proprietary or standards based file formats.

Encoding Images Synthesized from Light Field Image Data

[0084] An image synthesized from light field image data using super resolution processing can be encoded in accordance with the JPEG standard for inclusion in a light field image file in accordance with embodiments of the invention. The JPEG standard is a lossy compression standard. However, the information losses typically do not impact edges of objects. Therefore, the loss of information during the encoding of the image typically does not impact the accuracy of maps generated based upon the synthesized image (as opposed to the encoded synthesized image). The pixels within images contained within files that comply with the JFIF standard are typically encoded as YCbCr values. Many array cameras synthesize images, where each pixel is expressed in terms of a Red, Green and Blue intensity value. In several embodiments, the process of encoding the synthesized image involves mapping the pixels of the image from the RGB domain to the YCbCr domain prior to encoding. In other embodiments, mechanisms are used within the file to encode the image in the RGB domain. Typically, encoding in the YCbCr domain provides better compression ratios and encoding in the RGB domain provides higher decoded image quality.

Storing Additional Metadata Derived from Light Field Image Data

[0085] The JFIF standard does not specify a format for storing depth maps or auxiliary data generated by an array camera. The JFIF standard does, however, provide sixteen Application markers that can be utilized to store metadata concerning the encoded image contained within the file. In a number of embodiments, one or more of the Application markers of a JFIF file is utilized to store an encoded depth map and/or one or more auxiliary maps that can be utilized in the post processing of the encoded image contained within the file.

[0086] A JFIF Application marker segment that can be utilized to store a depth map, individual camera occlusion data and auxiliary map data in accordance with an embodiment of the invention is illustrated in FIG. 4. The APP9 Application marker segment 400 uses a format identifier 402 that uniquely identifies that the Application marker segment contains metadata describing an image synthesized using light field image data. In a number of embodiments, the identifier is referred to as the “DZ Format Identifier” 402 and is expressed as the zero terminated string “PIDZ0”.

[0087] The Application marker segment includes a header 404 indicated as “DZ Header” that provides a description of the metadata contained within the Application marker segment. In the illustrated embodiment, the “DZ Header” 404 includes a DZ Endian field that indicates whether the data in the “DZ Header” is big endian or little endian. The “DZ Header” 404 also includes a “DZ Selection Descriptor”.

[0088] An embodiment of a “DZ Selection Descriptor” is illustrated in FIG. 5, which includes four bytes. The first two bytes (i.e. bytes 0 and 1) contain information concerning the metadata describing the encoded image that are present (see FIG. 6) and the manner in which the different pieces of metadata are compressed (see FIG. 7). In the illustrated embodiment, the types of metadata that are supported are a depth map, occluded pixel data, virtual view point data, a missing pixel map, a regular edge map, a silhouette edge map, and/or a confidence map. In other embodiments, any of a variety of metadata describing an encoded image obtained from the light field image data used to synthesize the image can be included in the metadata contained within a JFIF file in accordance with an embodiment of the invention. In many instances, the metadata describing the encoded image can include maps that can be considered to be monochrome images that can be encoded using JPEG encoding. In a number of embodiments, the maps can be compressed using lossless JPEG LS encoding. In several embodiments, the maps can be compressed using lossy JPEG encoding. Utilizing JPEG encoding to compress the maps reduces the size of the maps and enables rendering devices to leverage a JPEG decoder to both decode the image contained within the JFIF file and the maps describing the encoded image. The third byte (i.e. byte 2) of the “DZ Selection Descriptor” indicates the number of sets of metadata describing the encoded image that are contained within the Application marker segment and the fourth byte is reserved. Although specific implementations of the header 404 describing the metadata contained within the Application marker segment are illustrated in FIGS. 4-7, any of a variety of implementations can be utilized to identify the maps describing the synthesized image that are present within the metadata contained within an light field image file as appropriate to the requirements of the application in accordance with embodiments of the invention.

Depth Map

[0089] Referring back to FIG. 4, the Application marker segment also includes a “Depth Map Header” 406 that describes depth map 416 included within the Application marker segment. The “Depth Map Header”406 includes an indication 408 of the size of “Depth Map Attributes” 410 included within the “Depth Map Header”, the “Depth Map Attributes” 410, and a “Depth Map Descriptor” 412. As noted above, the depth map 416 can be considered to be a monochrome image and lossless or lossy JPEG encoding can be utilized to compress the “Depth Map Data” included in a JFIF file.

[0090] A “Depth Map Attributes” table in accordance with an embodiment of the invention is illustrated in FIG. 8 and includes information concerning the manner in which the depth map should be used to render the encoded image. In the illustrated embodiment, the information contained within the “Depth Map Attributes” table includes the focal plane and the F# of the synthetic aperture to utilize when rendering the encoded image. Although specific pieces of information related to the manner in which the depth map can be utilized to render the encoded image are illustrated in FIG. 8, any of a variety of pieces of information appropriate to the requirements of a specific application can be utilized in accordance with embodiments of the invention.

[0091] A “Depth Map Descriptor” in accordance with an embodiment of the invention is illustrated in FIG. 9 and includes metadata describing the depth map. In the illustrated embodiment, the “Depth Map Descriptor” includes a zero terminated identifier string “PIDZDH0” and version information. In other embodiments, any of a variety of pieces of information appropriate to the specific requirements of particular applications can be utilized in accordance with embodiments of the invention.

[0092] A JFIF Application marker segment is restricted to 65,533 bytes. However, an Application marker can be utilized multiple times within a JFIF file. Therefore, depth maps in accordance with many embodiments of the invention can span multiple APP9 Application marker segments. The manner in which depth map data is stored within an Application marker segment in a JFIF file in accordance with an embodiment of the invention is illustrated in FIG. 10. In the illustrated embodiment, the depth map data is contained within a descriptor that is uniquely identified using the “PIDZDD0” zero terminated string. The descriptor also includes the length of the descriptor and depth map data.

[0093] Although specific implementations of a depth map and header describing a depth map within an Application marker segment of a JFIF file are illustrated in FIGS. 4, 8, 9, and 10, any of a variety of implementations can be utilized to include a depth map describing an encoded image within a JFIF file as appropriate to the requirements of the application in accordance with embodiments of the invention.

Occlusion Data

[0094] Referring back to FIG. 4, the Application marker segment also includes a “Camera Array Header” 418 that describes occlusion data 428 for individual cameras within an array camera that captured the light field image data utilized to synthesize the image contained within the light field image file. The occlusion data can be useful in a variety of post processing processes including (but not limited) to process that involve modifying the viewpoint of the encoded image. The “Camera Array Header” 418 includes an indication 420 of the size of a “Camera Array General Attributes” table 422 included within the “Camera Array Header”, the “Camera Array General Attributes” table 422, and a “Camera Array Descriptor” 424.

[0095] A “Camera Array General Attributes” table in accordance with an embodiment of the invention is illustrated in FIG. 11 and includes information describing the number of cameras and dimensions of a camera array utilized to capture the light field image data utilized to synthesize the image encoded within the JFIF file. In addition, the “Camera Array General Attributes” table can indicate a reference camera position within the array and/or a virtual view position within the array. The “Camera Array General Attributes” table also provides information concerning the number of cameras within the array for which occlusion data is provided within the JFIF file.

[0096] A “Camera Array Descriptor” in accordance with an embodiment of the invention is illustrated in FIG. 12 and includes metadata describing the individual camera occlusion data contained within the JFIF file. In the illustrated embodiment, the “Camera Array Descriptor” includes a zero terminated identifier string “PIDZAH0” and version information. In other embodiments, any of a variety of pieces of information appropriate to the specific requirements of particular applications can be utilized in accordance with embodiments of the invention.

[0097] In many embodiments, occlusion data is provided on a camera by camera basis. In several embodiments, the occlusion data is included within a JFIF file using an individual camera descriptor and an associated set of occlusion data. An individual camera descriptor that identifies a camera and identifies the number of occluded pixels related to the identified camera described within the JFIF file in accordance with an embodiment of the invention is illustrated in FIG. 13. In the illustrated embodiment, the descriptor is identified using the “PIDZCD0” zero terminated string. The descriptor also includes a camera number that can be utilized to identify a camera within an array camera that captured light field image data utilized to synthesize the encoded image contained within the JFIF file. In addition, the descriptor includes the number of occluded pixels described in the JFIF file and the length (in bytes) of the data describing the occluded pixels. The manner in which the occluded pixel data can be described in accordance with embodiments of the invention is illustrated in FIG. 14. The same descriptor “PDIZCD0” is used to identify the occluded pixel data and the descriptor also includes the number of pixels of occluded data contained within the segment, the length of the data in bytes and an offset to the next marker in addition to the occluded pixel data. Due to the restriction on Application marker segments not exceeding 65,533 bytes in data, the additional information enables a rendering device to reconstruct the occluded pixel data across multiple APP9 application marker segments within a JFIF file in accordance with embodiments of the invention.

[0098] A table describing an occluded pixel that can be inserted within a JFIF file in accordance with an embodiment of the invention is illustrated in FIG. 15. The table includes the depth of the occluded pixel, the pixel color of the occluded pixel and the pixel coordinates. In the illustrated embodiment, the pixel color is illustrated as being in the RGB domain. In other embodiments, the pixel color can be expressed in any domain including the YCbCr domain.

[0099] Although specific implementations for storing information describing occluded pixel depth within an Application marker segment of a JFIF file are illustrated in FIGS. 4, 13, 14, and 15, any of a variety of implementations can be utilized to include occluded pixel information within a JFIF file as appropriate to the requirements of the application in accordance with embodiments of the invention.

Auxiliary Maps

[0100] Referring back to FIG. 4, any of a variety of auxiliary maps can be included in an Application marker segment within a JFIF file in accordance with an embodiment of the invention. The total number of auxiliary maps and the types of auxiliary maps can be indicated in the Application marker segment. Each auxiliary map can be expressed using an “Auxiliary Map Descriptor” 432 and “Auxiliary Map Data” 434. In the illustrated embodiment, the “Auxiliary Map Descriptor” 432 is included in an “Auxiliary Map Header” 430 within the Application marker segment in the JFIF file.

[0101] An “Auxiliary Map Descriptor” that describes an auxiliary map contained within a light field image file in accordance with an embodiment of the invention is illustrated in FIG. 16. The “Auxiliary Map Descriptor” includes an identifier, which is the “PIDZAM0” zero terminated string and information specifying the type of auxiliary map and number of bits per pixel in the map. As noted above, any of a variety of auxiliary maps derived from light field image data used to synthesize an encoded image can be included within a JFIF file in accordance with embodiments of the invention. In the illustrated embodiment, confidence maps, silhouette edge maps, regular edge maps, and missing pixel maps are supported.

[0102] “Auxiliary Map Data” stored in a JFIF file in accordance with an embodiment of the invention is conceptually illustrated in FIG. 17. The “Auxiliary Map Data” uses the same “PDIZAD0” zero terminated string identifier and includes the number of pixels of the auxiliary map contained within the segment, the length of the data in bytes and an offset to the next marker in addition to pixels of the auxiliary map. Due to the restriction on Application marker segments not exceeding 65,533 bytes in data, the additional information enables a rendering device to reconstruct the auxiliary map describing the encoded image across multiple APP9 application marker segments within a JFIF file.

[0103] Although specific implementations for storing auxiliary maps within an Application marker segment of a JFIF file are illustrated in FIGS. 4, 16, and 17, any of a variety of implementations can be utilized to include auxiliary map information within a JFIF file as appropriate to the requirements of the application in accordance with embodiments of the invention. Various examples of auxiliary maps that can be utilized to provide additional information concerning an encoded image based upon the light field image data utilized to synthesize the encoded image in accordance with embodiments of the invention are discussed below.

Confidence Maps

[0104] A confidence map can be utilized to provide information concerning the relative reliability of the information at a specific pixel location. In several embodiments, a confidence map is represented as a complimentary one bit per pixel map representing pixels within the encoded image that were visible in only a subset of the images used to synthesize the encoded image. In other embodiments, a confidence map can utilize additional bits of information to express confidence using any of a variety of metrics including (but not limited to) a confidence measure determined during super resolution processing, or the number of images in which the pixel is visible.

Edge Maps

[0105] A variety of edge maps can be provided included (but not limited to) a regular edge map and a silhouette map. A regular edge map is a map that identifies pixels that are on an edge in the image, where the edge is an intensity discontinuity. A silhouette edge maps is a map that identifies pixels that are on an edge, where the edge involves an intensity discontinuity and a depth discontinuity. In several embodiments, each can be expressed as a separate one bit map or the two maps can be combined as a map including two pixels per map. The bits simply signal the presence of a particular type of edge at a specific location to post processing processes that apply filters including (but not limited to) various edge preserving and/or edge sharpening filters.

Missing Pixel Maps

[0106] A missing pixel map indicates pixel locations in a synthesized image that do not include a pixel from the light field image data, but instead include an interpolated pixel value. In several embodiments, a missing pixel map can be represented using a complimentary one bit per pixel map. The missing pixel map enables selection of post-processing filters to improve image quality. In many embodiments, a simple interpolation algorithm can be used during the synthesis of a higher resolution from light field image data and the missing pixels map can be utilized to apply a more computationally expensive interpolation process as a post processing process. In other embodiments, missing pixel maps can be utilized in any of a variety of different post processing process as appropriate to the requirements of a specific application in accordance with embodiments of the invention.

Rendering Images Using Light Field Imaging Files

[0107] When light field image data is encoded in a light field image file, the light field image file can be shared with a variety of rendering devices including but not limited to cameras, mobile devices, personal computers, tablet computers, network connected televisions, network connected game consoles, network connected media players, and any other device that is connected to the Internet and can be configured to display images. A system for sharing light field image files in accordance with an embodiment of the invention is illustrated in FIG. 18. The system 1800 includes a mobile device 1802 including an array camera configured to capture light field image data and encode the light field image data in a light field image file. The mobile device 1802 also includes a network interface that enables the transfer of a light field image file to other rendering devices via the Internet 1804. In several embodiments, the light field image file is transferred with the assistance of a server system 1806 that can either store the light field image file for access by other devices or relay the light field image file to other rendering devices. In many embodiments, the server system 1806 provides a user interface that enables users to modify the rendering of the image provided to the device. In several embodiments, the server system 1806 provides the light field image file to a device for rendering. In the illustrated embodiment, a variety of network connected rendering devices 1808 are illustrated including a mobile phone and a personal computer. In other embodiments, any of a variety of network connected and/or disconnected devices can render images using a light field image file in accordance with embodiments of the invention. Rendering devices and processes for rendering images in accordance with embodiments of the invention are discussed further below.

Rendering Devices

[0108] A rendering device in accordance with embodiments of the invention typically includes a processor and a rendering application that enables the rendering of an image based upon a light field image data file. The simplest rendering is for the rendering device to decode the encoded image contained within the light field image data file. More complex renderings involve applying post processing to the encoded image using the metadata contained within the light field image file to perform manipulations including (but not limited to) modifying the viewpoint of the image and/or modifying the focal plane of the image.

[0109] A rendering device in accordance with an embodiment of the invention is illustrated in FIG. 19. The rendering device 1900 includes a processor 1902, memory 1904, and an optional network interface 1906. The memory contains a rendering application 1908 that is used to configure the microprocessor to render images for display using a light field image file 1910. In the illustrated embodiment, the light field image file is shown stored in memory. In other embodiments, the light field image file can be stored in an external storage device. Although a specific rendering device is illustrated in FIG. 19, any of a variety of rendering devices can be utilized in accordance with embodiments of the invention including (but not limited to) the types of devices that are customarily used to display images using image files. Processes for rendering of images using light field image files in accordance with embodiments of the invention are discussed further below.

Processes for Rendering Images Using Light Field Image Files

[0110] As noted above, rendering a light field image file can be as simple as decoding an encoded image contained within the light field image file or can involve more complex post processing of the encoded image using metadata derived from the same light field image data used to synthesize the encoded image. A process for rendering a light field image in accordance with an embodiment of the invention is illustrated in FIG. 20. The process 2000 includes parsing (2002) the light field image file to locate the encoded image contained within the image file. The encoded image file is decoded (2004). As noted above, the image can be encoded using a standards based encoder and so the decoding process can utilize a standards based codec within a rendering device, or the image can be encoded using a proprietary encoding and a proprietary decoder is provided on the rendering device to decode the image. When the process for rendering the image simply involves rendering the image, the decoded image can be displayed. When the process for rendering the image includes post processing, the image file is parsed (2006) to locate metadata within the file that can be utilized to perform the post processing. The metadata is decoded (2008). The metadata can often take the form of maps that can be encoded using standards based image encoders and a standards based decoder present on the rendering device can be utilized to decode the metadata. In other embodiments, a proprietary decoding process is utilized to decode the metadata. The metadata can then be used to perform (2010) the post processing of the encoded image and the image can be displayed (2012). The display of the image can be local. Alternatively the image can be streamed to a remote device or encoded as an image and provided to a remote device for display.

[0111] Although specific processes for rendering an image from a light field image file are discussed with reference to FIG. 20, any of a variety of processes appropriate to the requirements of a specific application can be utilized to render an image for display using a light field image file in accordance with an embodiment of the invention. As noted above, any of a variety of standards based encoders and decoders can be utilized in the encoding and decoding of light field image files in accordance with embodiments of the invention. Processes for rendering images using light field image files that conform to the JFIF standard and include an image and/or metadata encoded in accordance with the JPEG standard are discussed further below.

Rendering Images Using JFIF Light Field Image Files

[0112] The ability to leverage deployed JPEG decoders can greatly simplify the process of rendering light field images. When a light field image file conforms to the JFIF standard and the image and/or metadata encoded within the light field image file is encoded in accordance with the JPEG standard, a rendering application can leverage an existing implementation of a JPEG decoder to render an image using the light field image file. Similar efficiencies can be obtained where the light field image file includes an image and/or metadata encoded in accordance with another popular standard for image encoding.

[0113] A rendering device configured by a rendering application to render an image using a light field image file in accordance with an embodiment of the invention is illustrated in FIG. 21. The rendering device 2100 includes a processor 2102, memory 2104, and an optional network interface 2106 that can be utilized to receive light field image files. In the illustrated embodiment, the memory 2104 of the rendering device 2100 includes a rendering application 2108, a JPEG decoder application 2110, and a light field image file 2112 that contains at least one image and/or metadata encoded in accordance with the JPEG standard. The rendering application 2108 configures the processor to parse the light field image file to locate an encoded image and to decode the encoded image using the JPEG decoder application 2110. Similarly, the rendering application can configure the processor to parse the light field image file to locate metadata and to decode encoded maps contained within the metadata using the JPEG decoder.

[0114] Although specific rendering devices including JPEG decoders are discussed above with reference to FIG. 21, any of a variety of rendering devices incorporating standards based decoders can be utilized to render images from appropriately encoded light field image files in accordance with embodiments of the invention. Processes for decoding light field image files that confirm with the JFIF standard and that contain at least one image and/or metadata encoded in accordance with the JPEG standard in accordance with embodiments of the invention are discussed further below.

Processes for Rendering Images From JFIF Light Field Image Files

[0115] Processes for rending images using light field image files that conform to the JFIF standard can utilize markers within the light field image file to identify encoded images and metadata. Headers within the metadata provide information concerning the metadata present in the file and can provide offset information or pointers to the location of additional metadata and/or markers within the file to assist with parsing the file. Once appropriate information is located a standard JPEG decoder implementation can be utilized to decode encoded images and/or maps within the file.

[0116] A process for displaying an image rendered using a light field image file that conforms to the JFIF standard using a JPEG decoder in accordance with an embodiment of the invention is illustrated in FIG. 22. The process 2200 involves parsing (2202) the light field image file to locate a Start of Image (SOI) Marker. The SOI marker is used to locate an image file encoded in accordance with the JPEG format. The encoded image can be decoded (2204) using a JPEG decoder. When no post processing of the decoded image is desired, the image can simply be displayed. Where post processing of the image is desired (e.g. to change the view point of the image and/or the focal plane of the image), the process parses (2206) the light field image file to locate an appropriate Application marker. In the illustrated embodiment, an APP9 marker indicates the presence of metadata within the light field image file. The specific metadata within the file can be determined by parsing (2206) a header within the APP9 Application marker segment that describes the metadata within the file. In the illustrated embodiment, the header is the “DZ Header” within the APP9 Application marker segment. The information within the metadata header can be utilized to locate (2208) specific metadata utilized in a post processing process within the light field image file. In instances where the metadata is encoded, the metadata can be decoded. In many embodiments, metadata describing an encoded image within a light field image file is in the form of a map that provides information concerning specific pixels within an encoded image contained within the light field image file and JPEG encoding is used to compress the map. Accordingly, a JPEG decoder can be utilized to decode the map. The decoded metadata can be utilized to perform (2212) a post processes the decoded image. The image can then be displayed (2214). In many embodiments, the image is displayed on a local display. In a number of embodiments, the image is streamed to a remote display or encoded as an image and forwarded to a remote device for display.

[0117] Although specific processes for displaying images rendered using light field image files are discussed above with respect to FIG. 22, any of a variety of processes for parsing a light field image file and decoding images and/or metadata encoded in accordance with the JPEG standard using a JPEG decoder can be utilized in accordance with embodiments of the invention. Much of the discussion above references the use of metadata derived from light field image data and contained within a light field image file to perform post processing processes on an encoded image synthesized from the light field image data. Post processing of images synthesized from light field image data using metadata obtained using the light field image data in accordance with embodiments of the invention are discussed further below.

Post Processing of Images Using Metadata Derived from Light Field Image Data

[0118] Images can be synthesized from light field image data in a variety of ways. Metadata included in light field image files in accordance with embodiments of the invention can enable images to be rendered from a single image synthesized from the light field image data without the need to perform super resolution processing. Advantages of rendering images in this way can include that the process of obtaining the final image is less processor intensive and less data is used to obtain the final image. However, the light field image data provides rich information concerning a captured scene from multiple viewpoints. In many embodiments, a depth map and occluded pixels from the light field image data (i.e. pixels that are not visible from the reference viewpoint of the synthesized image) can be included in a light field image file to provide some of the additional information typically contained within light field image data. The depth map can be utilized to modify the focal plane when rendering an image and/or to apply depth dependent effects to the rendered image. The depth map and the occluded pixels can be utilized to synthesize images from different viewpoints. In several embodiments, additional maps are provided (such as, but not limited to, confidence maps, edge maps, and missing pixel maps) that can be utilized when rendering alternative viewpoints to improve the resulting rendered image. The ability to render images from different viewpoints can be utilized to simply render an image from a different viewpoint. In many embodiments, the ability to render images from different viewpoints can be utilized to generate a stereo pair for 3D viewing. In several embodiments, processes similar to those described in U.S. Provisional Patent Application Ser. No. 61/707,691, entitled “Synthesizing Images From Light Fields Utilizing Virtual Viewpoints” to Jain (the disclosure of which is incorporated herein by reference in its entirety) can be utilized to modify the viewpoint based upon motion of a rendering device to create a motion parallax effect. Processes for rendering images using depth based effects and for rendering images using different viewpoints are discussed further below.

Rendering Images Using Depth Based Effects

[0119] A variety of depth based effects can be applied to an image synthesized from light field image data in accordance with embodiments of the invention including (but not limited to) applying dynamic refocusing of an image, locally varying the depth of field within an image, selecting multiple in focus areas at different depths, and/or applying one or more depth related blur model. A process for applying depth based effects to an image synthesized from light field image data and contained within a light field image file that includes a depth map in accordance with an embodiment of the invention is illustrated in FIG. 23. The process 2300 includes decoding (2302) an image synthesized from light field image data contained within a light field image file. In addition, a depth map derived from the light field image data that describes the synthesized image is also decoded (2304) from metadata contained within the light field image file. One or more depth dependent effects can then be applied (2406) to the pixels of the decoded image based upon the depths of the pixels indicated by the depth map. In a number of embodiments, the depth dependent effects are automatically determined by modifying the focal plane, and/or F number (which provides different depths of fields and degree of blur in out-of-focus regions). The image can then be displayed (2308). In many embodiments, the image is displayed on a local display. In a number of embodiments, the image is streamed to a remote display or encoded as an image and forwarded to a remote device for display.

[0120] Although specific processes for applying depth dependent effects to an image synthesized from light field image data using a depth map obtained using the light field image data are discussed above with respect to FIG. 23, any of a variety of processes can be utilized for extracting an image and a depth map from a light field image file and for using the depth map to apply one or more depth dependent effects in accordance with embodiments of the invention. Processes for rendering images from different viewpoints to the reference viewpoint of an image contained within a light field image file based upon a depth map and information concerning occluded pixels contained within the light field image file in accordance with embodiments of the invention are discussed further below.

Rendering Images Using Different Viewpoints

[0121] One of the compelling aspects of computational imaging is the ability to use light field image data to synthesize images from different viewpoints. The ability to synthesize images from different viewpoints creates interesting possibilities including the creation of stereo pairs for 3D applications and the simulation of motion parallax as a user interacts with an image. Light field image files in accordance with many embodiments of the invention can include an image synthesized from light field image data from a reference viewpoint, a depth map for the synthesized image and information concerning pixels from the light field image data that are occluded in the reference viewpoint. A rendering device can use the information concerning the depths of the pixels in the synthesized image and the depths of the occluded images to determine the appropriate shifts to apply to the pixels to shift them to the locations in which they would appear from a different viewpoint. Occluded pixels from the different viewpoint can be identified and locations on the grid of the different viewpoint that are missing pixels can be identified and hole filling can be performed using interpolation of adjacent non-occluded pixels. In many embodiments, the quality of an image rendered from a different viewpoint can be increased by providing additional information in the form of auxiliary maps that can be used to refine the rendering process. In a number of embodiments, auxiliary maps can include confidence maps, edge maps, and missing pixel maps. Each of these maps can provide a rendering process with information concerning how to render an image based on customized preferences provided by a user. In other embodiments, any of a variety of auxiliary information including additional auxiliary maps can be provided as appropriate to the requirements of a specific rendering process.

[0122] A process for rendering an image from a different viewpoint using a light field image file containing an image synthesized using light field image data from a reference viewpoint, a depth map describing the depth of the pixels of the synthesized image, and information concerning occluded pixels in accordance with an embodiment of the invention is illustrated in FIG. 24. The process 2400 includes decoding (2402) an image contained within a light field image file, where the image is an image synthesized from light field image data. The process also includes decoding (2404) a depth map from the light field image file, where the depth map was also obtained from the light field image data used to synthesize the encoded image. Information concerning pixels from the light field image data that are occluded in the reference viewpoint is also obtained (2405) from the light field image file, where the information includes the location and depth of the occluded pixels from the reference viewpoint. In many embodiments, auxiliary information, including auxiliary maps that specify additional information concerning the pixels in the encoded image, is also contained within the light field image file and auxiliary information useful in the rendering of an image from a different viewpoint to the reference viewpoint can be extracted and decoded (2408) from the light field image file. Using the depth map and the depths of the occluded pixels, shifts in the location and depths of pixels in the different viewpoint can be determined (2410). Based upon the shifts, occluded pixels can be determined (2414) and the image displayed. Where auxiliary information is available, the auxiliary information can be utilized to adjust (2412) the pixels in the image prior to rendering. In many embodiments, the adjustments are performed prior to identifying occluded pixels and displaying the final image. In a number of embodiments, the adjustments are performed after occluded pixels are identifies.

[0123] Although specific processes for rendering an image from a different viewpoint using an image synthesized from a reference view point using light field image data, a depth map obtained using the light field image data, and information concerning pixels in the light field image data that are occluded in the reference viewpoint are discussed above with respect to FIG. 24, any of a variety of processes can be utilized for rendering images from different viewpoints using a light field image file as appropriate to the requirements of specific applications in accordance with embodiments of the invention. Processes for rendering images simulating different lens characteristics from in accordance with embodiments of the invention are discussed further below.

[0124] While the above description contains many specific embodiments of the invention, these should not be construed as limitations on the scope of the invention, but rather as an example of one embodiment thereof. Accordingly, the scope of the invention should be determined not by the embodiments illustrated, but by the appended claims and their equivalents.