A method of depth map coding for a three-dimensional video coding system incorporating consistent texture merging candidate is disclosed. According to the first embodiment, the current depth block will only inherit the motion information of the collocated texture block if one reference depth picture has the same POC (picture order count) and ViewId (view identifier) as the reference texture picture of the collocated texture block. In another embodiment, the encoder assigns the same total number of reference pictures for both the depth component and the collocated texture component for each reference list. Furthermore, the POC (picture order count) and the ViewId (view identifier) for both the depth image unit and the texture image unit are assigned to be the same for each reference list and for each reference picture.

A method of depth map coding for a three-dimensional video coding system incorporating consistent texture merging candidate is disclosed. According to the first embodiment, the current depth block will only inherit the motion information of the collocated texture block if one reference depth picture has the same POC (picture order count) and ViewId (view identifier) as the reference texture picture of the collocated texture block. In another embodiment, the encoder assigns the same total number of reference pictures for both the depth component and the collocated texture component for each reference list. Furthermore, the POC (picture order count) and the ViewId (view identifier) for both the depth image unit and the texture image unit are assigned to be the same for each reference list and for each reference picture.

A system receives an encoded image representative of the 2D projection of a cubic image, the encoded image generated from two overlapping hemispherical images separated along a longitudinal plane of a sphere. The system decodes the encoded image to produce a decoded 2D projection of the cubic image, and perform a stitching operation to portions of the decoded 2D projection representative of overlapping portions of the hemispherical images to produce stitched overlapping portions. The system combine the stitched overlapping portions with portions of the decoded 2D projection representative of the non-overlapping portions of the hemispherical images to produce a stitched 2D projection of the cubic image, and encode the stitched 2D projection of the cubic image to produce an encoded cubic projection of the stitched hemispherical images.

A system receives an encoded image representative of the 2D projection of a cubic image, the encoded image generated from two overlapping hemispherical images separated along a longitudinal plane of a sphere. The system decodes the encoded image to produce a decoded 2D projection of the cubic image, and perform a stitching operation to portions of the decoded 2D projection representative of overlapping portions of the hemispherical images to produce stitched overlapping portions. The system combine the stitched overlapping portions with portions of the decoded 2D projection representative of the non-overlapping portions of the hemispherical images to produce a stitched 2D projection of the cubic image, and encode the stitched 2D projection of the cubic image to produce an encoded cubic projection of the stitched hemispherical images.

An image encoding method includes, for each of left and right images of an image stereo pair, identifying a region of interest (ROI) in an input image, generating multiple representations of that input image such that one representation is a high field of view (FOV) representation including substantially all of the input image and another representation is a low FOV representation having a lower FOV than the high FOV representation and including at least the ROI. The relationship between pixel resolution and image size for the high and low FOV representations is such that, in relation to the ROI, the low FOV representation provides more image detail than the high FOV representation. The method also includes generating one or more multiplexed images by juxtaposing portions of the left and right, high and low FOV representations so as to include all image content thereof in fewer than four multiplexed images.

An image encoding method includes, for each of left and right images of an image stereo pair, identifying a region of interest (ROI) in an input image, generating multiple representations of that input image such that one representation is a high field of view (FOV) representation including substantially all of the input image and another representation is a low FOV representation having a lower FOV than the high FOV representation and including at least the ROI. The relationship between pixel resolution and image size for the high and low FOV representations is such that, in relation to the ROI, the low FOV representation provides more image detail than the high FOV representation. The method also includes generating one or more multiplexed images by juxtaposing portions of the left and right, high and low FOV representations so as to include all image content thereof in fewer than four multiplexed images.

An encoding device evaluates a plurality of processing and/or post-processing algorithms and/or methods to be applied to a video stream, and signals a selected method, algorithm, class or category of methods/algorithms either in an encoded bitstream or as side information related to the encoded bitstream. A decoding device or post-processor utilizes the signaled algorithm or selects an algorithm/method based on the signaled method or algorithm. The selection is based, for example, on availability of the algorithm/method at the decoder/post-processor and/or cost of implementation. The video stream may comprise, for example, downsampled multiplexed stereoscopic images and the selected algorithm may include any of upconversion and/or error correction techniques that contribute to a restoration of the downsampled images.

The disclosed subject matter is directed to employing machine learning models configured to predict 3D data from 2D images using deep learning techniques to derive 3D data for the 2D images. In some embodiments, a method is provided that comprises receiving, by a system operatively coupled to a processor, a two-dimensional image, and determining, by the system, auxiliary data for the two-dimensional image, wherein the auxiliary data comprises orientation information regarding a capture orientation of the two-dimensional image. The method further comprises, deriving, by the system, three-dimensional information for the two-dimensional image using one or more neural network models configured to infer the three-dimensional information based on the two-dimensional image and the auxiliary data

The disclosed subject matter is directed to employing machine learning models configured to predict 3D data from 2D images using deep learning techniques to derive 3D data for the 2D images. In some embodiments, a method is provided that comprises receiving, by a system operatively coupled to a processor, a two-dimensional image, and determining, by the system, auxiliary data for the two-dimensional image, wherein the auxiliary data comprises orientation information regarding a capture orientation of the two-dimensional image. The method further comprises, deriving, by the system, three-dimensional information for the two-dimensional image using one or more neural network models configured to infer the three-dimensional information based on the two-dimensional image and the auxiliary data

An encoding device evaluates a plurality of processing and/or post-processing algorithms and/or methods to be applied to a video stream, and signals a selected method, algorithm, class or category of methods/algorithms either in an encoded bitstream or as side information related to the encoded bitstream. A decoding device or post-processor utilizes the signaled algorithm or selects an algorithm/method based on the signaled method or algorithm. The selection is based, for example, on availability of the algorithm/method at the decoder/post-processor and/or cost of implementation. The video stream may comprise, for example, downsampled multiplexed stereoscopic images and the selected algorithm may include any of upconversion and/or error correction techniques that contribute to a restoration of the downsampled images.