Abstract: Various schemes pertaining to pre-encoding processing of a video stream with motion compensated temporal filtering (MCTF) are described. An apparatus determines a filtering interval for a received raw video stream having pictures in a temporal sequence. The apparatus selects from the pictures a plurality of target pictures based on the filtering interval, as well as a group of reference pictures for each target picture to perform pixel-based MCTF, which generates a corresponding filtered picture for each target picture. The apparatus subsequently transmits the filtered pictures as well as non-target pictures to an encoder for encoding the video stream. Subpictures of natural images and screen content images are separately processed by the apparatus.

Abstract: A semiconductor device includes a plurality of channel layers vertically separated from one another. The semiconductor device also includes an active gate structure comprising a lower portion and an upper portion. The lower portion wraps around each of the plurality of channel layers. The semiconductor device further includes a gate spacer extending along a sidewall of the upper portion of the active gate structure. The gate spacer has a bottom surface. Moreover, a dummy gate dielectric layer is disposed between the gate spacer and a topmost channel layer of plurality of channel layers. The dummy gate dielectric layer is in contact with a top surface of the topmost channel layer, the bottom surface of the gate spacer, and the sidewall of the gate structure.

Abstract: For each prediction candidate of a set of one or more prediction candidates of the current block, a video coder computes a matching cost between a set of reference pixels of the prediction candidate in a reference picture and a set of neighboring pixels of a current block in a current picture. The video coder identifies a subset of the reference pictures as major reference pictures based on a distribution of the prediction candidates among the reference pictures of the current picture. A bounding block is defined for each major reference picture, the bounding block encompassing at least portions of multiple sets of reference pixels for multiple prediction candidates. The video coder assigns an index to each prediction candidate based on the computed matching cost of the set of prediction candidates. A selection of a prediction candidate is signaled by using the assigned index of the selected prediction candidate.

Abstract: For each prediction candidate of a set of one or more prediction candidates of the current block, a video coder computes a matching cost between a set of reference pixels of the prediction candidate in a reference picture and a set of neighboring pixels of a current block in a current picture. The video coder identifies a subset of the reference pictures as major reference pictures based on a distribution of the prediction candidates among the reference pictures of the current picture. A bounding block is defined for each major reference picture, the bounding block encompassing at least portions of multiple sets of reference pixels for multiple prediction candidates. The video coder assigns an index to each prediction candidate based on the computed matching cost of the set of prediction candidates. A selection of a prediction candidate is signaled by using the assigned index of the selected prediction candidate.

Abstract: Video processing methods and apparatuses for processing a current block in a current picture include receiving input data of the current block, determining a reference picture, determining whether picture sizes of the current and reference pictures are different, determining whether horizontal wraparound motion compensation is enabled for predicting the current block, performing motion compensation for the current block to obtain a reference block from the reference picture, and encoding or decoding the current block according to the reference block. Horizontal wraparound motion compensation is disabled when the picture sizes of the current and reference pictures are different.

Abstract: Method and apparatus of coding VR360 video sequence are disclosed, wherein wraparound motion compensation is included as a coding tool. According to the method, a bitstream corresponding to encoded data of the VR360 video sequence is generated at an encoder side or received at a decoder side, where the bitstream comprises one or more PPS syntaxes related to wraparound motion compensation information in a PPS (Picture Parameter Set). The VR360 video sequence is encoded at the encoder side or decoded at the decoder side based on the wraparound motion compensation information.

Abstract: An electronic apparatus and a compression method for an artificial neural network are provided. The compression method is adapted for the artificial neural network with a plurality of convolution layers. The compression method includes: setting a first pruning layer for coupling the first pruning layer to Lth convolution layer, where the first pruning layer has a plurality of first weighting values and each of the first weighting values corresponds to each of a plurality of channels of the Lth convolution layer; tuning the first weighting values, selecting a part of the channels of the Lth convolution layer to be at least one first redundancy channel according to the first weighting values, and generating a compressed Lth convolution layer by deleting the at least one first redundancy channel; and removing the first pruning layer, and generating a first compressed artificial neural network.

Abstract: Video processing methods and apparatuses include receiving input video data, determining a Merge number for a regular Merge mode, comparing the Merge number for the regular Merge mode with a predefined value, and only signaling or parsing a difference between the Merge number for the regular Merge mode and a Merge number for a Geometric Partition Mode (GPM) Merge mode when the Merge number for the regular Merge mode is larger than or equal to the predefined value. The difference between the Merge numbers is inferred to 0 when the difference is not signaled. The Merge number for the GPM Merge mode defines a size of a GPM Merge candidate list constructed for each block coded or to be coded by the GPM Merge mode.

Abstract: Video processing methods and apparatuses include receiving input video data, determining a Merge number for a regular Merge mode, comparing the Merge number for the regular Merge mode with a predefined value, and only signaling or parsing a difference between the Merge number for the regular Merge mode and a Merge number for a Geometric Partition Mode (GPM) Merge mode when the Merge number for the regular Merge mode is larger than or equal to the predefined value. The difference between the Merge numbers is inferred to 0 when the difference is not signaled. The Merge number for the GPM Merge mode defines a size of a GPM Merge candidate list constructed for each block coded or to be coded by the GPM Merge mode.

Abstract: Video processing methods and apparatuses for processing a current block in a current picture include receiving input data of the current block, determining a reference picture, determining whether picture sizes of the current and reference pictures are different, determining whether horizontal wraparound motion compensation is enabled for predicting the current block, performing motion compensation for the current block to obtain a reference block from the reference picture, and encoding or decoding the current block according to the reference block. Horizontal wraparound motion compensation is disabled when the picture sizes of the current and reference pictures are different.

Abstract: An electronic apparatus and a compression method for an artificial neural network are provided. The compression method is adapted for the artificial neural network with a plurality of convolution layers. The compression method includes: setting a first pruning layer for coupling the first pruning layer to Lth convolution layer, where the first pruning layer has a plurality of first weighting values and each of the first weighting values corresponds to each of a plurality of channels of the Lth convolution layer; tuning the first weighting values, selecting a part of the channels of the Lth convolution layer to be at least one first redundancy channel according to the first weighting values, and generating a compressed Lth convolution layer by deleting the at least one first redundancy channel; and removing the first pruning layer, and generating a first compressed artificial neural network.