Google Inc.

Qunshan Gu

Wei Jia

Yaowu Xu

Marco Paniconi

Ronghua Wu

Paul Gordon Wilkins

Xun Cai

Object-based intra-prediction encoding may include generating, by a processor in response to instructions stored on a non-transitory computer readable medium, an encoded block of a current frame of a video stream by encoding a current block from the current frame, including the encoded block in an output bitstream, and outputting or storing the output bitstream. Encoding the current block may include identifying a first spatial portion of the current block, wherein the first spatial portion includes a first pixel from the current block and omits a second pixel from the current block, encoding the first pixel using a first intra-prediction mode, and encoding the second pixel using a second intra-prediction mode, wherein the second intra-prediction mode differs from the first intra-prediction mode.

Object-based intra-prediction

Encoding is accomplished using a reference picture list. Reference frames are identified with picture numbers and are assigned to reference picture buffers when used for encoding. The picture numbers associated the reference frames in the reference picture buffers are ranked. Reordering can occur to assign short word lengths to those reference frames used frequently for prediction. The resulting encoded reference picture list is sent to a decoder. The decoder decodes the reference picture list. By comparing picture numbers in the reference picture list with picture numbers of reference picture buffers of the decoder, the encoder and decoder buffers can remain in sync even under packet loss conditions. For example, the decoder can use the decoded reference picture list to determine missing and unused reference frames.

Video data encoding using reference picture lists

Encoding and decoding is accomplished herein using a multi-reference picture buffer. Decoding includes receiving an encoded video stream, determining an expected error value representing possible frame errors, decoding a plurality of reference frames from the video stream, adding each of the plurality of reference frames to the reference picture buffers up to a maximum number of reference picture buffers, and determining an encoded video stream error value representing actual frame errors based on the decoded plurality of reference frames. If the encoded video stream error value is greater than the decoded expected error value, an error is signaled.

Managing multi-reference picture buffers for video data coding

A method for encoding a frame having a plurality of blocks in a video stream using transform-domain intra prediction is disclosed. The method includes generating, using a two-dimensional transform, a set of transform coefficients for a current block; generating, using a one-dimensional transform, a set of transform coefficients for previously coded pixel values in the frame; determining, using the set of transform coefficients for the previously coded pixel values, a set of transform coefficients for a prediction block; determining a residual based on the difference between the set of transform coefficients for the current block and the set of coefficients for the prediction block; and encoding the residual.

Transform-domain intra prediction

Techniques for efficient coding of digital media data are presented. A resource controller component dynamically allocates computing resources between an estimator component and a coder component. The estimator component generates an initial motion estimation of a raw video frame of a sequence of raw video frames based on a previous raw video frame. The coder component encodes the previous raw video frame to generate a reconstructed video frame in parallel with the generation of the initial motion estimation. When the previous raw video frame is reconstructed, the coder component dynamically determines whether and/or how to refine the motion estimation to encode the raw video frame based on a predefined encoding criterion. In making that determination, the coder component selects and uses a mode(s) of motion estimation refinement from a plurality of available motion estimation refinement modes to encode the raw video frame.

Encoding time management in parallel real-time video encoding

Edge-selective intra coding may include encoding a first input block, generating a first reconstructed block by decoding the first encoded block, determining a reconstruction error block based on a difference between the first input block and the first reconstructed block, identifying a second input block from the first input frame, and determining a plurality of refined context pixels. Determining a plurality of refined context pixels may include identifying a plurality of context pixels from the reconstruction error block, generating a plurality of encoded context pixels by encoding the plurality of context pixels, generating a plurality of reconstructed context pixels by decoding the plurality of encoded context pixels, and adding at least a portion of the reconstruction error block and the plurality of reconstructed context pixels. The method may include generating a second encoded block by encoding the second input block based on the plurality of refined context pixels.

Edge-selective intra coding

Coding a video stream having a plurality of frames is disclosed. A frame of a video stream is processed to determine if direct motion prediction can be applied to the macroblocks of the frame. Macroblocks to which direct motion can be applied are grouped into superblocks. The distribution of superblock size and configuration is analyzed to determine variable-length codes that can vary for each frame or segment encoded using direct prediction superblocks, thereby reducing the number of bits used to indicate superblock encoding in the encoded video stream.

Variable-sized super block based direct prediction mode

Object-based intra-prediction decoding may include generating a reconstructed block by decoding an encoded block, which may include identifying a first intra-prediction mode for a first portion of the reconstructed block and identifying a second intra-prediction mode for a second portion of the reconstructed block. Decoding the encoded block may include generating a first portion of a prediction block using the first intra-prediction mode, generating a second portion of the prediction block using the second intra-prediction mode, generating the first portion of the reconstructed block based on the first portion of the prediction block, such that the first portion of the reconstructed block corresponds to a first object represented in the reconstructed block, and generating the second portion of the reconstructed block based on the second portion of the prediction block, such that the second portion of the reconstructed block corresponds to a second object represented in the reconstructed block.

Second-order orthogonal spatial intra-prediction may include generating a reconstructed frame by generating a plurality of reconstructed blocks by decoding a portion of the current encoded frame, generating a decoded residual block by decoding a current block from the current encoded frame, identifying a current pixel location in the current block, identifying reconstructed pixels in respective reconstructed blocks, wherein first and second reconstructed pixels are at respective oblique angle to the current pixel location, and a third reconstructed pixel at the second oblique angle to the first reconstructed pixel or at the first oblique angle to the second reconstructed pixel, determining a predictor pixel for the current pixel location using a difference between the third reconstructed pixel and a sum of the first reconstructed pixel and the second reconstructed pixel, and identifying a reconstructed pixel as a sum of the predictor pixel and a corresponding residual pixel.

Second-order orthogonal spatial intra prediction

Encoding and decoding is accomplished herein using a reference picture list. Reference frames are identified with picture numbers and are assigned to reference picture buffers when used for encoding. The picture numbers associated the reference frames in the reference picture buffers are ranked. Reordering can occur to assign short word lengths to those reference frames used frequently for prediction. The resulting encoded reference picture list is sent to a decoder. The decoder decodes the reference picture list. By comparing picture numbers in the reference picture list with picture numbers of reference picture buffers of the decoder, the encoder and decoder buffers can remain in sync even under packet loss conditions. For example, the decoder can use the decoded reference picture list to determine missing and unused reference frames.

Video data encoding and decoding using reference picture lists

Conditional predictive multi-symbol run-length coding may include generating a differential residual block based on a residual block and a reconstructed residual block, entropy encoding the differential residual block by determining a symbol order for a plurality of symbols in the differential residual block, grouping the plurality of symbols into at least a first group and a second group based on the symbol order, run-length encoding the plurality of symbols in the differential residual block, by on a condition that a first symbol from the plurality of symbols is associated with the first group, encoding a second symbol from the plurality of symbols, and on a condition that the first symbol is associated with the second group, modifying the second symbol and encoding the modified second symbol.

Conditional predictive multi-symbol run-length coding

Encoding and decoding a video image having a plurality of frames using a two-step quantization and coding process are disclosed. A block of a frame are encoded by identifying pixels having certain spatial characteristics, forming a second block from the block while replacing the identified pixels with a single pixel value, such as an average of the remaining original pixels. The second block is encoded, such as by transformation and quantization, and placed into a bitstream. The second block is decoded and subtracted from the original block to generate a difference block. The difference block is encoded, such as by quantization, and is placed in the bitstream. At a decoder, both blocks are decoded and combined to reconstruct the original block.

Two-step quantization and coding method and apparatus

Encoding and decoding is accomplished herein using a multi-reference picture buffer. Reference frames are assigned to reference picture buffers when used for encoding and are identified with picture IDs unique relative to other picture IDs currently associated with the reference picture buffers. The maximum picture number used as the picture ID can be based on a number of reference picture buffers available to encode and decode the frames plus a value based on an expected error rate. The picture IDs can be assigned based on a least recently used policy. When a reference frame is no longer needed for encoding and decoding, a picture ID number associated with the reference frame can be released for re-assignment to a new reference frame.

Managing multi-reference picture buffers and identifiers to facilitate video data coding

A block of a frame of a video stream can be encoded using lossless coding that generates a transform domain residual block and a spatial domain difference block. The compression ratio of the coding may be improved by selecting an optimal quantization value on a per-block basis. The optimal quantization value can be selected by quantizing a residual block resulting from prediction of the block using a plurality of candidate quantization values and selecting the candidate quantization value that results in the fewest number of bits for the quantized residual block.

Lossless video coding with sub-frame level optimal quantization values

A method and apparatus for performing object-based intra-prediction encoding and decoding is disclosed. Implementations of object-based intra-prediction encoding and decoding include segmenting a current block into a first portion and at least a second portion; generating, for at least some of a plurality of pixels in the first portion, corresponding predicted values according to at least a first intra-prediction mode; generating, for at least some of a plurality of pixels in the second portion, corresponding predicted values according to at least a second intra-prediction mode; and encoding the current block using the predicted values of the first portion and the predicted values of the second portion.

Growing community of inventors

Hayward, CA, United States of America

Qunshan Gu