Abstract: Methods are described for determining thresholds to skip entropy coding of latent features in image and video coding using a neural network. A threshold based on a mean of estimated standard deviations of all latents is proposed. For autoregressive neural networks, to avoid drift between context model parameter estimates computed during training and inference, it is proposed to have two separate entropy parameter estimation networks: an initial entropy-estimation network, to be used in entropy skip encoding and decoding of quantized latents, and a refined entropy-estimation network, to be used in arithmetic encoding and decoding of the quantized latents. Methods and systems for multi-stage entropy skipping are also proposed.

Abstract: Methods, systems, bitstream syntax, and fixed-point implementations are described for loop filtering using neural networks in image and video processing. Given an input image, a hybrid luma-chroma filter is proposed, wherein luma and chroma components are first processed by a first neural network and the output of the first network is subsequently processed by separate luma and chroma subnetworks. Finally, the outputs of the separate luma and chroma subnetworks are concatenated to generate the filtered output of the input image. Computational efficient methods using CP-decomposition are also described. Methods indicating the position of the neural-net loop filter relatively to other filters, such as the adaptive loop filter (ALF) are also discussed.

Abstract: Methods and systems are described for video coding and decoding using intra affine prediction. Two or three control point best vectors (CPBV) to be used to derive motion vectors of affine predictions are generated based on creating for each CPBV a list of candidate best vectors, and selecting according to a criterion the two or three best ones. Methods to generate the list of the candidate best vectors include a first method based on neighboring coded units and a distance criterion among two consecutive best vectors in the list, and a second method based on template matching between a coded unit to be encoded or decoded and prior-decoded coded units in a reference area.

Abstract: Methods and systems are described for video coding and decoding using cross-component prediction tools. The proposed methods include: inter-luma based modeling for chroma, cross component residue prediction, and convolutional cross-component intra prediction.

Abstract: Methods and systems are described for intra-prediction using template matching (TM) in video coding. The proposed methods include adaptive fusion when using template-based intra mode derivation using the most probable modes (TIMD), and fusion in intra mode prediction with template matching (Intra TMP).

Abstract: Methods, systems, and bitstream syntax are described for the fusion of latent features in multi-level, end-to-end, neural networks used in image and video compression. The fused architecture may be static or dynamic based on image characteristics (e.g., natural images versus screen content images) or other coding parameters, such as bitrate constrains or rate-distortion optimization. A variety of multi-level fusion architectures are discussed.

Abstract: Methods and systems are described for intra-prediction using template matching (TM) in video coding. The proposed methods include adaptive fusion when using template-based intra mode derivation using the most probable modes (TIMD), and fusion in intra mode prediction with template matching (Intra TMP).

Abstract: Methods, systems, and bitstream syntax are described for inter-frame coding using end-to-end neural networks used in image and video compression. Inter-frame coding methods include one or more of: joint luma-chroma motion compensation for YUV pictures, joint luma-chroma residual coding for YUV pictures, using attention layers, enabling temporal motion prediction networks for motion vector prediction, using a cross-domain network which combines motion vector and residue information for motion vectors decoding, using the cross-domain network for decoding residuals, using weighted motion-compensated inter prediction, and using temporal only, spatial only, or both temporal and spatial features in entropy decoding. Methods to improve training of neural networks for inter-frame coding are also described.

Abstract: Methods are described for template matching (TM) in video coding. The proposed methods include: the use of constrained top and left neighbors in template matching, enabling TM only in coding tree unit boundaries, using approximated reconstructed samples, a new processing pipeline for deriving decoder side intra mode derivation (DIMD) combined with template based intra mode derivation (TIMD), and using filtered pixels from the neighbors, instead of using the reconstructed pixels. Furthermore, methods are described on how template matching may be applied in combination with Intra, sub-partitioning mode, interpolation filtering in intra prediction, block partitioning, bi-prediction with coding unit-level weights, and adaptive motion vector resolution.

Abstract: Methods related to the geometric partition mode (GPM) in video coding are described. The proposed methods include: applying adaptive ordering of merge candidates with template matching (ARMC-TM) to derive GPM inter candidate lists, candidates list applying merge motion vector differences in GPM, enabling GPM for all-intra coding units (CUs), using inter and intra template costs in intra-prediction modes of GPM, using GPM partitions to generate templates in template matching, and using neighboring reconstructed samples and an edge criterion to derive top and left-edge intercepts to generate partitioning candidates.

Abstract: Methods, systems, and bitstream syntax are described for sign prediction in video coding. The method include: selection of top and left neighbors based on an image continuity check, the intra mode of the current coded unit (CU), the merge motion vector, or adaptive motion vector prediction, sign prediction based on residue domain of current CU or neighbor CUs, sign prediction based on approximated reconstruction samples, reducing the number of selected coefficients for sorting, simplifying the sequential search cost, and by combining sign prediction with sign data hiding.

Abstract: Methods, systems, and bitstream syntax are described for the entropy modeling of latent features in image and video coding using a combination of probability density functions. Using high-level syntax elements, an encoder may signal to compliant decoders the multi-distribution entropy model using: the number of one or more PDFs being used, an identifier of each PDF being used among a list of available PDFs, the number of PDF parameters in each PDF, and syntax elements indicating which PDF parameters across two or more PDFs being used are being shared.

Abstract: Methods, systems, and bitstream syntax are described for the fusion of latent features in multi-level, end-to-end, neural networks used in image and video compression. The fused architecture may be static or dynamic based on image characteristics (e.g., natural images versus screen content images) or other coding parameters, such as bitrate constrains or rate-distortion optimization. A variety of multi-level fusion architectures are discussed.

Abstract: Methods, systems, and bitstream syntax are described for the entropy modeling of latent features in image and video coding using a combination of probability density functions. Using high-level syntax elements, an encoder may signal to compliant decoders the multi-distribution entropy model using: the number of one or more PDFs being used, an identifier of each PDF being used among a list of available PDFs, the number of PDF parameters in each PDF, and syntax elements indicating which PDF parameters across two or more PDFs being used are being shared.

Abstract: Methods, processes, and systems are presented for combining signal reshaping (also referred to as luma mapping chroma residuals scaling) with local illumination compensation (LIC) in video coding. Examples and trade-offs when the LIC model parameters are computed in the original domain, the reshaped domain, or a mixed domain, are presented.

Abstract: Methods, processes, and systems are presented for combining signal reshaping (also referred to as luma mapping chroma residuals scaling) with local illumination compensation (LIC) in video coding. Examples and trade-offs when the LIC model parameters are computed in the original domain, the reshaped domain, or a mixed domain, are presented.

Abstract: Methods, processes, and systems are presented for combining signal reshaping (also referred to as luma mapping chroma residuals scaling) with local illumination compensation (LIC) in video coding. Examples and trade-offs when the LIC model parameters are computed in the original domain, the reshaped domain, or a mixed domain, are presented.

Abstract: Methods, processes, and systems are presented for combining signal reshaping (also referred to as luma mapping chroma residuals scaling) with local illumination compensation (LIC) in video coding. Examples and trade-offs when the LIC model parameters are computed in the original domain, the reshaped domain, or a mixed domain, are presented.