Interdigital Ce Patent Holdings, Sas

Interdigital Madison Patent Holdings, Sas

Thierry Dumas

Franck Galpin

Philippe Bordes

Tangi Poirier

Fabrice Le Leannec

Pavel Nikitin

Fabrice Urban

Ya Chen

Karam Naser

Fabrice LeLeannec

Gagan Bihari Rath

Methods and apparatuses for encoding/decoding a video are provided. A picture from a video is decoded from a bit-stream, the picture being encoded in the bitstream at a resolution that is lower than an original resolution of the picture. Decoding the picture further comprises up-sampling the decoded picture to the original resolution, and applying an adaptive filter to at least one component of the up-sampled decoded picture. According to some embodiments, the adaptive filter may be performed in-loop or out-of-loop of the decoding/encoding process.

Methods and apparatuses for encoding/decoding a video

Decoder side intra mode derivation (DIMD) capability is enhanced by using reference pixels that extend several rows above and several columns left of a current video block, and also includes pixels above and left of the current video block, as well as columns above and right and rows below and left of the current video block. The reference pixels are formed from surrounding reconstructed/previously encoded samples surrounding the current video block. The derivation of an intra prediction mode is determined from gradients determined from each of the reference pixels in a defined surrounding area. In one embodiment, the gradients are determined using horizontal and vertical filters. In a sub-embodiment, the filters do not extend beyond the defined surrounding area. In another embodiment, reassignment of an index indicative of the target intra prediction mode is performed.

Improving the angle discretization in decoder side intra mode derivation

At least a method and an apparatus are presented for efficiently encoding or decoding video. For example, an intra prediction of an image block using at least one neural network from a context comprising pixels surrounding the image block is determined and an information relative to a transform method to apply for decoding the image block is also determined. The transform method is adapted to the neural network intra prediction mode of the block to encode or decode. The information relative to the transform method is inferred from the at least one neural network used in intra prediction of the image block at the encoding and either signaled or also inferred at the decoding.

Adapting the transform process to neural network-based intra prediction mode

A method for controlling a display of a frame representing a state of a game in a network-based gaining application comprising: sending (A) an information representative of a user action to a server; receiving (B) at least one frame representing a predicted state of the game predicted from the user action; obtaining an information allowing determining if said at least one frame corresponds to the user action; and; if the at least one frame corresponds to said user action, determining when displaying one of the at least one frame in function of a time at which this user action was performed and of said information.

Display control in cloud gaming applications

A video coding system is provided that performs intra prediction in a mode using a neural network for block of only a set of specific block sizes. The signaling of this mode is designed to be efficient in terms of rate-distortion under this constraint. Different transformations of the context of a block and the neural network prediction of this block are introduced in order to use one single neural network for predicting blocks of several sizes, as well as the corresponding signaling. The neural network-based prediction mode considers both luminance blocks and chrominance blocks. The video coding system comprises encoder and decoder apparatuses, encoding, decoding and signal generation methods and a signal carrying information corresponding to the described coding mode.

Neural network-based intra prediction for video encoding or decoding

There is provided intra predicting of an image block using at least one neural network from a context comprising pixel surrounding said image block. Said intra predicting also uses an information to distinguish the missing pixels from the available pixels in the context. There are also provided encoding/decoding methods and apparatus, a device, a non-transitory computer readable medium and a computer program product.

Deep intra prediction of an image block

A method and an apparatus for deep prediction refinement are disclosed. A first motion-compensated region for a block of a picture and a second region for said block are obtained. A prediction for said block is determined using a Neural Network that uses said first motion-compensated region and said second region.

Deep prediction refinement

In one implementation, KLT transform matrices are derived during the encoding or decoding process. In particular, the KLT transform matrices can be derived for either the primary transforming stage or the secondary transforming stage, or both. Because the KLT transform matrices can be derived at both the encoder and decoder sides, the KLT transform matrices do not need to be signaled in the bitstream. To derive the KLT transform matrix for a current block to be encoded or decoded, a template for the current block is defined, and is used to search in reconstructed regions multiple blocks with templates similar to the current template. Those multiple blocks are used to train the KLT transform matrix.

Karhunen loeve transform for video coding

In one implementation, to perform in-loop filtering of a version of reconstructed samples of a block, only a single offset parameter is signaled in the bitstream. Based on the version of reconstructed samples, a pixel-wise weight mask is generated using a neural network. Because the neural network parameters are known at both the encoder and decoder, these parameters need not to be signaled in the bitstream. The single offset parameter scaled by the weighted mask is used to adjust the samples in the block. Thus, even though only a single offset parameter is used, the samples are adjusted by pixel-wise offsets. The neural network may also take other parameters, such as quantization parameters and picture types as input. Further, there can be multiple neural networks that generate different weight masks, where different offsets are signaled and one or more of the neural networks are to be selected for filtering.

Neural network based filter in video coding

The disclosure relates to a method for encoding image data, the method including intra-predicting, or predicting by combining inter-prediction and intra-prediction, a first block of the image data by using an intra-prediction mode using a first single transformation obtained by taking account of the first block size. The disclosure also relates to a method for encoding image data, a variable coding length being used for signaling a plurality of Prediction Modes by the encoding, the method comprising:—intra-predicting a first block of the image data by using an intra-prediction mode using a first transformation, the first transformation being obtained by taking account of the first block size, —encoding information signaling a use of the intra-prediction mode in a bitstream, the information being encoded as one of the plurality of Prediction Modes. The disclosure further relates to the corresponding decoding methods, devices and media.

Growing community of inventors

Rennes, France

Thierry Dumas