Disclosed is a method of decoding an image and a method of encoding an image. The method of decoding an image includes: obtaining motion-constrained tile set information; determining, on the basis of the motion-constrained tile set information, a first boundary region of a collocated tile set within a reference picture, which corresponds to a motion-constrained tile set; padding a second boundary region corresponding to the first boundary region; and performing inter prediction on the motion-constrained tile set by using a collocated tile set that includes the padded second boundary region.

An example method of video processing includes making a first determination, for a conversion between a current block of a video and a bitstream representation of the video, about whether a sample in a neighboring block of the current block is in a same video region as the current block and using the first determination to make a second determination about applicability of a coding tool that uses samples outside the current block to the conversion of the current block. The coding tool comprises an adaptive loop filter (ALF) tool that comprises an ALF classification process and/or an ALF filtering process. The method also includes performing the conversion according to the first determination and the second determination.

The present disclosure provides systems and methods for video coding. The systems include, for example, an image encoder comprising: circuitry; and a memory coupled to the circuitry, wherein the circuitry, in operation, performs the following: predicting a first block of prediction samples for a current block of a picture, wherein predicting the first block of prediction samples includes at least a prediction process with a motion vector from a different picture; padding the first block of prediction samples to form a second block of prediction samples, wherein the second block is larger than the first block; calculating at least a gradient using the second block of prediction samples; and encoding the current block using at least the calculated gradient.

The present disclosure provides systems and methods for video coding. The systems include, for example, an image encoder comprising: circuitry; and a memory coupled to the circuitry, wherein the circuitry, in operation, performs the following: predicting a first block of prediction samples for a current block of a picture, wherein predicting the first block of prediction samples includes at least a prediction process with a motion vector from a different picture; padding the first block of prediction samples to form a second block of prediction samples, wherein the second block is larger than the first block; calculating at least a gradient using the second block of prediction samples; and encoding the current block using at least the calculated gradient.

In a case of referring to information between pictures by means of inter prediction, the entire picture needs to be decoded even in a case of decoding a small region. In a case of referring to information within a picture by means of intra prediction, referring to information across a tile boundary is not possible. In a case that a pixel pointed by a sub-block level motion vector of a target block that is calculated by scaling based on an available motion vector acquired from a spatial neighboring block or a temporal neighboring block is not present within a tile sequence, a process of replacing the pixel value with a pixel value within the tile sequence is performed. In a case of referring to a pixel outside of a tile by means of intra prediction, a process of replacing the pixel value with that of a pixel within the tile is performed.

In a case of referring to information between pictures by means of inter prediction, the entire picture needs to be decoded even in a case of decoding a small region. In a case of referring to information within a picture by means of intra prediction, referring to information across a tile boundary is not possible. In a case that a pixel pointed by a sub-block level motion vector of a target block that is calculated by scaling based on an available motion vector acquired from a spatial neighboring block or a temporal neighboring block is not present within a tile sequence, a process of replacing the pixel value with a pixel value within the tile sequence is performed. In a case of referring to a pixel outside of a tile by means of intra prediction, a process of replacing the pixel value with that of a pixel within the tile is performed.

Methods and apparatus for using flexible grid regions in picture or video frames are disclosed. In one embodiment, a method includes receiving a set of first parameters that defines a plurality of first grid regions comprising a frame. For each first grid region, the method includes receiving a set of second parameters that defines a plurality of second grid regions, and the plurality of second grid regions partitions the respective first grid region. The method further includes partitioning the frame into the plurality of first grid regions based on the set of first parameters, and partitioning each first grid region into the plurality of second grid regions based on the respective set of second parameters.

Methods and apparatus for using flexible grid regions in picture or video frames are disclosed. In one embodiment, a method includes receiving a set of first parameters that defines a plurality of first grid regions comprising a frame. For each first grid region, the method includes receiving a set of second parameters that defines a plurality of second grid regions, and the plurality of second grid regions partitions the respective first grid region. The method further includes partitioning the frame into the plurality of first grid regions based on the set of first parameters, and partitioning each first grid region into the plurality of second grid regions based on the respective set of second parameters.

Systems and methods are described for reducing the complexity of using bi-directional optical flow (BIO) in video coding. In some embodiments, bit-width reduction steps are introduced in the BIO motion refinement process to reduce the maximum bit-width used for BIO calculations. In some embodiments, simplified interpolation filters are used to generate predicted samples in an extended region around a current coding unit. In some embodiments, different interpolation filters are used for vertical versus horizontal interpolation. In some embodiments, BIO is disabled for coding units with small heights and/or for coding units that are predicted using a sub-block level inter prediction technique, such as advanced temporal motion vector prediction (ATMVP) or affine prediction.

Training a video decoder system may include masking one of at least two sets of video encoding parameters with invalid values to generate an invalid set. The at least two sets of video encoding parameters are provided to one or more neural networks to train them to predict valid values that correspond to values of the invalid set using an iterative training algorithm. Encoding parameters to encode are determined based on a prediction error of the one or more neural networks. Encoding parameters which are determined to be accurately predicted are dropped from the encoded data. A new video stream is encoded without the dropped encoding parameters.