In one embodiment, an apparatus comprises a memory and a processor. The memory is to store visual data associated with a visual representation captured by one or more sensors. The processor is to: obtain the visual data associated with the visual representation captured by the one or more sensors, wherein the visual data comprises uncompressed visual data or compressed visual data; process the visual data using a convolutional neural network (CNN), wherein the CNN comprises a plurality of layers, wherein the plurality of layers comprises a plurality of filters, and wherein the plurality of filters comprises one or more pixel-domain filters to perform processing associated with uncompressed data and one or more compressed-domain filters to perform processing associated with compressed data; and classify the visual data based on an output of the CNN.

Disclosed herein are an image encoding method and an image decoding method. The image decoding method includes determining an initial motion vector of a current block using a motion vector of a reconstructed region, searching for the motion vector of the current block based on the initial motion vector, and generating a prediction sample of the current block using the motion vector. The initial motion vector includes a motion vector in a past direction and a motion vector in a future direction.

Provided is an image decoding method performed by a decoding apparatus including deriving a temporary reference motion vector for a current block included in a current picture, deriving a reference motion vector for the current block by applying clipping to the temporary reference motion vector, deriving a motion vector for the current block based on the reference motion vector, deriving a predicted block for the current block based on the motion vector for the current block, and generating a reconstructed picture for the current picture based on the predicted block for the current block, wherein a number of bits of a binary code representing the reference motion vector is less than a number of bits of a binary code representing the temporary reference motion vector.

An image decoding method performed by a decoding apparatus according to the present disclosure includes receiving a bitstream including residual information; deriving quantized transform coefficients for a current block based on the residual information included in the bitstream; deriving residual samples for the current block based on the quantized transform coefficients; and generating a reconstructed picture based on the residual samples for the current block.

The disclosure discloses a method for processing video signals and an apparatus therefor. Specifically, the method of processing video signals based on inter prediction, comprising: configuring a merge candidate list based on a spatial neighboring block and a temporal neighboring block to a current block; adding a history-based merge candidate of the current block to the merge candidate list; obtaining a merge index indicating a merge candidate used for an inter prediction of the current block within the merge candidate list; generating a prediction sample of the current block based on motion information of the merge candidate indicated by the merge index; and updating a history-based merge candidate list based on the motion information.

The disclosure discloses a method for processing video signals and an apparatus therefor. Specifically, the method of processing video signals based on inter prediction, comprising: configuring a merge candidate list based on a spatial neighboring block and a temporal neighboring block to a current block; adding a history-based merge candidate of the current block to the merge candidate list; obtaining a merge index indicating a merge candidate used for an inter prediction of the current block within the merge candidate list; generating a prediction sample of the current block based on motion information of the merge candidate indicated by the merge index; and updating a history-based merge candidate list based on the motion information.

A method for video encoding includes setting a motion vector precision associated with a current block to be encoded in a current picture, and determining a motion vector for encoding the current block based on the motion vector precision. The method also includes determining a motion vector difference for the current block based on (i) the determined motion vector for encoding the current block, (ii) a predicted motion vector of the current block in inter prediction mode, and (iii) the motion vector precision. The method further includes encoding the current block according to the determined motion vector, and generating a coded video bitstream including the encoded current block and including prediction information indicating that the current block is coded in inter prediction mode and indicating the determined motion vector difference for the current block.

A method for video encoding includes setting a motion vector precision associated with a current block to be encoded in a current picture, and determining a motion vector for encoding the current block based on the motion vector precision. The method also includes determining a motion vector difference for the current block based on (i) the determined motion vector for encoding the current block, (ii) a predicted motion vector of the current block in inter prediction mode, and (iii) the motion vector precision. The method further includes encoding the current block according to the determined motion vector, and generating a coded video bitstream including the encoded current block and including prediction information indicating that the current block is coded in inter prediction mode and indicating the determined motion vector difference for the current block.

A video encoder or decoder processes portions of video using virtual temporal affine motion candidates. Under the general aspects, virtual temporal affine candidates are created using only the classical temporal motion buffer information, avoiding the storage of additional affine parameters in a temporal motion buffer. A motion field for encoding or decoding a video block is generated based on the virtual temporal affine candidates. In one embodiment, collocated motion candidates are rescaled by adjusting the picture order count of the determined motion field. In another embodiment, resolution adaptation is performed to enable a current motion buffer to correspond to a reference motion buffer.

A method for video encoding includes determining a corresponding motion vector for each of multiple control points of a base predictor. The method further includes determining a corresponding motion vector difference for each of the multiple control points of the base predictor based on the determined motion vector for each respective control point. The method further includes generating prediction information of the current block to be included in a coded video bitstream. The prediction information includes (i) a usage flag indicative of the affine merge mode with offset, (ii) offset parameters defining the determined corresponding motion vector difference for each of the one or more of the control points, and (iii) a zero motion vector difference flag for the multiple control points of the base predictor. The zero motion vector difference flag indicates whether offset parameters for the respective control point are provided in the prediction information.