Google Inc.

On2 Technologies, Inc.

Adrian Grange

Paul Gordon Wilkins

Yaowu Xu

James Bankoski

Eric Ameres

Tim Murphy

Francis William Galligan

Dan Miller

Scott Lavarnway

Ronald Sebastiaan Bultje

Jingning Han

Ching-Han Chiang

Yuxin Liu

Jim Bankoski

Alexander Jay Converse

Jamieson Brettle

Stanislav Vitvitskyy

Scott J La Varnway

John Koleszar

Video coding may include identifying an input frame from an input video stream, generating, by a processor, an output bitstream by encoding the input frame, wherein encoding the input frame includes dividing the input frame into at least a first contiguous area of the input frame and a second contiguous area of the input frame, generating first encoded data by encoding the first contiguous area of the input frame, generating second encoded data by encoding the second contiguous area of the input frame, wherein encoding the second contiguous area of the input frame is performed independently of encoding the first contiguous area of the input frame, and including the first encoded data and the second encoded data in the output bitstream, and outputting the output bitstream.

Independently coding frame areas

Decoding a video sequence may include generating decoded video data by decoding encoded video data from an encoded bitstream, wherein decoding the encoded video data includes identifying candidate reference motion vectors used to decode previously decoded blocks, identifying reconstructed pixel values from a block spatially adjacent to a current block in the current frame, for each candidate reference motion vector, determining a respective score based on a difference between the reconstructed pixel values and pixel values obtained using the candidate reference motion vector, identifying a candidate reference motion vector having the minimal score from the candidate reference motion vectors as the reference motion vector, decoding a motion vector from the encoded bitstream using the reference motion vector, and decoding the current block from the encoded bitstream using the motion vector.

Video coding using reference motion vectors

Regions for texture-based coding are identified using a spatial segmentation and a motion flow segmentation. For frames of a group of frames in a video sequence, a frame is segmented using a first classifier into at least one of a texture region or a non-texture region of an image in the frame. Then, the texture regions of the group of frames are segmented using a second classifier into a texture coding region or a non-texture coding region. The second classifier uses motion across the group of frames as input. Each of the classifiers is generated using a machine-learning process. Blocks of the non-texture region and the non-texture coding region of the current frame are coded using a block-based coding technique, while blocks of the texture coding region are coded using a coding technique that is other than the block-based coding technique.

Video compression through motion warping using learning-based motion segmentation

A system, apparatus, and method for decoding a video signal having a plurality of frames including a current frame. Decoding the video signal includes generating an alternate reference frame, detecting an error in the current frame and reporting the detected error, receiving a recovery frame in response to reporting the detected error, wherein the recovery frame is encoded using the alternate reference frame; and decoding the recovery frame using the alternate reference frame. The alternate reference frame occurs prior to the current frame in the video signal and is other than a frame immediately prior to the current frame.

Error recovery using alternate reference frame

A first interpolation filter is selected for a first direction of the prediction block, and a second interpolation filter is selected for a second direction of the prediction block. The first interpolation filter has a first number of taps and the second interpolation filter has a second number of taps. The second interpolation filter is set to an interpolation filter having a third number of taps in response to determining that the first number of taps is greater than a threshold number of taps and the second number of taps is greater than a threshold number of taps. The third number of taps is smaller than or equal to the threshold number of taps. The prediction block is generated using the first interpolation filter and the second interpolation filter.

Single direction long interpolation filter

Techniques described use a reference motion vector to reduce the amount of bits needed to encode motion vectors for inter prediction. One method includes determining a reference motion vector used to encode a current block of a current frame. The reference motion vector was selected by identifying, multiple candidate motion vectors used to inter predict blocks, and identifying a set of reconstructed pixel values from at least one block adjacent to the current block. For each candidate motion vector, a set of predicted pixel values for the set of reconstructed pixel value is generated using the candidate motion vector and a first reference frame, and an error value is calculated based on a difference between the sets of pixel values. Finally, a reference motion vector from the candidates is selected based on the error values. The method also includes decoding the current block using the reference motion vector.

Systems and methods are disclosed for adaptation of presentation speed for content presentation (e.g., audio content presentation). For example, methods may include obtaining an indication of motion in a space; adjusting a presentation speed based on the indication of motion; and presenting a content item associated with a location in the space, wherein the content item is presented using the adjusted presentation speed.

Adaptation of presentation speed

A method for encoding a spherical video is disclosed. The method includes mapping a frame of the spherical video to a two dimensional representation based on a projection. Further, in a prediction process the method includes determine whether at least one block associated with a prediction scheme is on a boundary of the two dimensional representation, and upon determining the at least one block associated with the prediction scheme is on the boundary, select an adjacent end block as a block including at least one pixel for use during the prediction process, the adjacent end block being associated with two or more boundaries of the two dimensional representation.

Continuous prediction domain

Disclosed herein are methods and apparatuses for selective prediction signal filtering. One aspect of the disclosed implementations is a method for encoding a frame of a video stream including determining a first performance measurement for a first set of prediction samples identified for a group of pixels using a first prediction mode, generating a filtered set of prediction samples by applying a filter to a second set of prediction samples, wherein at least one of the filtered set of prediction samples or the second set of prediction samples are identified using a second prediction mode, determining a second performance measurement for the filtered set of prediction samples, generating, using a processor, a residual based on the filtered set of prediction samples and the group of pixels if the second performance measurement exceeds the first performance measurement, and encoding the frame using the residual.

Selective prediction signal filtering

Encoding and decoding using advance coded reference prediction may include identifying a sequence of temporally adjacent frames from the plurality of frames, wherein each frame in the sequence of temporally adjacent frames is associated with a respective frame position indicating a temporal location the sequence, encoding a first frame from the sequence as an intra-coded frame, generating an alternate reference frame by reconstructing the first encoded frame, encoding a second frame from the sequence with reference to a reference frame, the second frame associated with a second frame position, including the first encoded frame in a compressed bitstream at a first bitstream position, and including the second encoded frame in the compressed bitstream at a second bitstream position, wherein the second bitstream position is later than the first bitstream position and wherein the first frame position is later than the second frame position.

Advance coded reference prediction

A method for encoding a video stream having a plurality of blocks including a current block. The method includes identifying, from a plurality of previously coded blocks, a plurality of candidate motion vectors; identifying, for the current block, a set of reconstructed pixel values corresponding to a set of previously coded pixel values; determining, for at least one of the plurality of candidate motion vectors, an error value based on the set of reconstructed pixel values and a set of predicted values for the set of previously coded pixel values; and selecting, from the plurality of candidate motion vectors, a reference motion vector for the current block based on the error value.

Determine reference motion vectors

Techniques are described to use a reference motion vector to reduce the amount of bits needed to encode motion vectors for inter prediction. One method includes identifying a candidate motion vector used to inter predict each of a plurality of previously coded blocks to define a plurality of candidate motion vectors, identifying a set of reconstructed pixel values corresponding to a set of previously coded pixels for the current block, and generating, using each candidate motion vector, a corresponding set of predicted values for the set of previously coded pixel values within each reference frame of a plurality of reference frames. A respective error value based on a difference between the set of reconstructed pixel values and each set of predicted values is used to select a reference motion vector from the candidate motion vectors that is used to encode the motion vector for the current block.

A current block and a reference frame used to compress the current block are selected from a bitstream. The reference frame has an alternate reference block created by obtaining a preliminary alternate reference block for the current block, obtaining a predictor block that is generated based on a frame other than the current frame, and creating the alternate reference block when the difference between the two predictors is less than a threshold. The alternate reference block is created by using certain blocks from the sequence of frames that are more different from the preliminary alternate reference block than other non-selected blocks. The current block is decoded using the alternate reference block. In this way, a predictor block that is different from a golden frame predictor is made available when an existing alternate reference block is too similar to the golden frame predictor to be useful.

Apparatus and method for using an alternate reference frame to decode a video frame

Implementations of the teachings herein include coding video data with an alternate reference frame generated using a temporal filter. The alternate reference frame is generated by determining a first weighting factor, for each corresponding block of a respective frame of a filter set, that represents a temporal correlation of the block with the corresponding block, determining a second weighting factor, for each pixel for each corresponding block of the respective frame of the filter set, that represents a temporal correlation of the pixel to a spatially-correspondent pixel in the block, determining a filter weight for each pixel in the block and for each spatially-correspondent pixel is each corresponding block based on the first weighting factor and the second weighting factor, and generating a weighted average pixel value for each pixel position in the block to form a block of the alternate reference frame based on the filter weights.

Growing community of inventors

Cambridgeshire, United Kingdom

Adrian Grange