Disclosed are a method for decoding a video signal and a device therefore. More particularly, a method for decoding an image based on an intra prediction mode, may comprise the steps of: acquiring a most probable mode (MPM) flag for indicating whether an MPM is applied to a current block, the MPM representing a mode in which an intra prediction mode of the current block is induced from an intra predicted block neighboring the current block; configuring an MPM list, on the basis of whether intra prediction modes of a left neighboring block and an above neighboring block of the current block satisfy predefined conditions, if the MPM is applied to the current block; acquiring an MPM index for indicating, within the MPM list, a specific intra prediction mode; and generating a prediction block for the current block by using the intra prediction mode specified by the MPM index.

An image decoding method of decoding, on a block-by-block basis, image data included in a coded stream includes: deriving candidates for an intra prediction mode to be used for intra prediction for a decoding target block, the number of the candidates constantly being a plural number; obtaining, from the coded stream, an index for identifying one of the derived candidates for the intra prediction mode; and determining, based on the obtained index, one of the derived candidates for the intra prediction mode as the intra prediction mode to be used for intra prediction for the decoding target block.

Method For Implementing A Quantizer In A Multimedia Compression And Encoding System is disclosed. In the Quantizer system of the present invention, several new quantization ideas are disclosed. In one embodiment, adjacent macroblocks are grouped together into macroblock groups. The macroblock groups are then assigned a common quantizer value. The common quantizer value may be selected based upon how the macroblocks are encoded, the type of macroblocks within the macroblock group (intra-blocks or inter-blocks), the history of the motion vectors associated with the macroblocks in the macroblock group, the residuals of the macroblocks in the macroblock group, and the energy of the macroblocks in the macroblock group. The quantizer value may be adjusted in a manner that is dependent on the current quantizer value. Specifically, if the quantizer value is at the low end of the quantizer scale, then only small adjustments are made. If the quantizer value is at the high end then larger adjustments may be made to the quantizer. Finally, in one embodiment, the quantizer is implemented along with an inverse quantizer for efficient operation.

A method and apparatus for video encoding and decoding with prediction offset in a video coding system are disclosed. A prediction offset for a current block is derived from a reconstructed residual signal of one or more coded blocks. At the encoder side, a residual signal for the current block is calculated by subtracting a prediction signal and the prediction offset from an original signal of the current block, and the residual signal is encoded to generate an encoded bitstream. At the decoder side, a residual signal for the current block is obtained from input data, a reconstructed signal for the current block is calculated by adding a prediction signal and the prediction offset to the residual signal, and the current block is decoded using the reconstructed signal.

A method of decoding JVET video, comprising defining a coding unit (CU) template within a decoded area of a video frame, the CU template being positioned above and/or to the left of a current decoding position for which data was intra predicted, defining a search window within the decoded area, the search window being adjacent to the CU template, generating a plurality of candidate prediction templates based on pixel values in the search window, each of the plurality of candidate prediction templates being generated using different intra prediction modes, calculating a matching cost between the CU template and each of the plurality of candidate prediction templates, selecting an intra prediction mode that generated the candidate prediction template that had the lowest matching cost relative to the CU template, and generating a prediction CU for the current decoding position based on the intra prediction mode.

The technology described herein relates to online per-title encoding. A method for online per-title encoding includes receiving a video input, generating segments of the video input, extracting a spatial feature and a temporal feature, predicting bitrate-resolution pairs based on the spatial feature and the temporal feature, using a discrete cosine transform (DCT)-based energy function, and per-title encoding segments of the video input for the predicted bitrate-resolution pairs. A system for online per-title encoding may include memory for storing a set of bitrates, a set of resolutions, and a machine learning module configured to predict bitrate resolution pairs based on low-complexity spatial and temporal features.

An image coding apparatus configured to divide an image into one or more slices each including a plurality of blocks and to code each slice on a block-by-block basis includes a first coding unit configured to code blocks included in a first portion of the slice, and a second coding unit configured to code blocks included in a second portion of the slice, wherein, when the second coding unit codes an initial block in the second portion, the second coding unit codes the initial included in the second portion by referring to a first quantization parameter provided to the slice as an initial value and referred to by the first coding unit when the first coding units codes the initial block in the first portion.

A machine learning (ML) task system trains a neural network model that learns a compressed representation of acquired data and performs a ML task using the compressed representation. The neural network model is trained to generate a compressed representation that balances the objectives of achieving a target codelength and achieving a high accuracy of the output of the performed ML task. During deployment, an encoder portion and a task portion of the neural network model are separately deployed. A first system acquires data, applies the encoder portion to generate a compressed representation, performs an encoding process to generate compressed codes, and transmits the compressed codes. A second system regenerates the compressed representation from the compressed codes and applies the task model to determine the output of a ML task.

Disclosed are a method for decoding a video signal and a device therefore. More particularly, a method for decoding an image based on an intra prediction mode, may comprise the steps of: acquiring a most probable mode (MPM) flag for indicating whether an MPM is applied to a current block, the MPM representing a mode in which an intra prediction mode of the current block is induced from an intra predicted block neighboring the current block; configuring an MPM list, on the basis of whether intra prediction modes of a left neighboring block and an above neighboring block of the current block satisfy predefined conditions, if the MPM is applied to the current block; acquiring an MPM index for indicating, within the MPM list, a specific intra prediction mode; and generating a prediction block for the current block by using the intra prediction mode specified by the MPM index.

Methods and apparatus are provided for determining quantization parameter predictors from a plurality of neighboring quantization parameters. An apparatus includes an encoder for encoding image data for at least a portion of a picture using a quantization parameter predictor for a current quantization parameter to be applied to the image data. The quantization parameter predictor is determined using multiple quantization parameters from previously coded neighboring portions. A difference between the current quantization parameter and the quantization parameter predictor is encoded for signaling to a corresponding decoder.