Qualcomm Incorporated

Reza Pourreza

Taco Sebastiaan Cohen

Amir Said

Yang Yang

Yinhao Zhu

Hoang Cong Minh Le

Yizhe Zhang

Auke Joris Wiggers

Ties Jehan Van Rozendaal

Guillaume Konrad Sautiere

Mingu Lee

Roland Memisevic

Yadong Lu

Pulkit Madan

Adam Waldemar Golinski

Sunny Praful Kumar Panchal

Apratim Bhattacharyya

A processor-implemented method for video compression using an artificial neural network (ANN) includes receiving a video via the ANN. The ANN extracts a first set of features of a current frame of the video and a second set of features of a reference frame of the video. The ANN determines an estimate of correlation features between the first set of features of the current frame and the second set of features of the reference frame. The estimate of the correlation features are encoded and transmitted to a receiver.

Flow-agnostic neural video compression

A processor-implemented method for generating grounded rationales for visual reasoning tasks includes receiving, by a first artificial neural network (ANN), an interleaved sequence of images and textual information. The first ANN extracts grid features of the images of the interleaved sequence of the images and the textual information to generate a representation of the interleaved sequence of the images and the textual information based on the grid features. A second ANN maps the grid features to a textual domain. The second ANN extracts visual information of the interleaved sequence of the images and the textual information based on the grid features in the textual domain. The second ANN determines a rationale based on the visual information. The visual information comprises one or more lower-level surrogate tasks.

Using grounded rationales to improve visual reasoning

Systems and techniques are provided for coding (e.g., encoding and/or decoding) video data using camera motion information. For example, a decoding device can obtain a frame of encoded video data associated with an input frame, the frame of encoded video data including camera information associated with generating the video data and a residual. A camera motion compensated frame can be generated based on a reference frame and the camera information. Optical flow information associated with object motion determined based on at least the input frame and the reference frame can be generated. A motion compensated frame can be generated by warping the camera motion compensated frame based on the optical flow information. A reconstructed input frame can be generated based on the motion compensated frame and the residual.

Video coding using camera motion compensation and object motion compensation

Techniques are described for processing video data, such as by performing learned bidirectional coding using a unidirectional coding system and an interpolated reference frame. For example, a process can include obtaining a first reference frame and a second reference frame. The process can include generating a third reference frame at least in part by performing interpolation between the first reference frame and the second reference frame. The process can include performing unidirectional inter-prediction on an input frame based on the third reference frame, such as by estimating motion between an input frame and the third reference frame, and generating a warped frame at least in part by warping one or more pixels of the third reference frame based on the estimated motion. The process can include generating, based on the warped frame and a predicted residual, a reconstructed frame representing the input frame, the reconstructed frame including a bidirectionally-predicted frame.

Learned B-frame coding using P-frame coding system

Systems and techniques are provided for coding video data based on an optical flow correction and a residual correction. For example, a decoding device can obtain a frame of encoded video data associated with an input frame, the frame of encoded video data including an optical flow correction and a residual correction. A predicted optical flow can be generated based on one or more reference frames and a reference optical flow. A corrected prediction frame can be generated based on the predicted optical flow and the optical flow correction. A predicted residual can be generated based on at least the corrected prediction frame and a first reference frame included in the one or more reference frames. The decoding device can generate a reconstructed input frame based on the corrected prediction frame, the predicted residual, and the residual correction.

Video coding using optical flow and residual predictors

A processor-implemented method for image compression using an artificial neural network (ANN) includes receiving, at an encoder of the ANN, an image and a spatial segmentation map corresponding to the image. The spatial segmentation map indicates one or more regions of interest. The encoder compresses the image according to a controllable spatial bit allocation. The controllable spatial bit allocation is based on a learned quantization bin size.

Neural image compression with controllable spatial bit allocation

A computer-implemented method for operating an artificial neural network (ANN) includes receiving an input by the ANN. The ANN generates a latent representation of the input. The latent representation is communicated according to a bit rate based on a learned latent scaling parameter. The latent scaling parameter is learned based on a channel index and a tradeoff parameter value that corresponds to a value that balances the bit rate and a distortion.

Variable bit rate compression using neural network models

A method of image compression includes receiving an image. Multiple quantized latent representations are generated to represent features of the image. Each of the quantized latent representations has a different resolution and is generated at staggered timings. Each of the later generated quantized latent representations is conditioned on each of the prior generated quantized latent representations. The multiple quantized latent representations are decoded to reconstruct the image.

Data compression with a multi-scale autoencoder

Systems and techniques are described for encoding and/or decoding data based on motion estimation that applies variable-scale warping. An encoding device can receive an input frame and a reference frame that depict a scene at different times. The encoding device can generate an optical flow identifying movements in the scene between the two frames. The encoding device can generate a weight map identifying how finely or coarsely the reference frame can be warped for input frame prediction. The encoding device can generate encoded video data based on the optical flow and the weight map. A decoding device can generate a reconstructed optical flow and a reconstructed weight map from the encoded data. A decoding device can generate a prediction frame by warping the reference frame based on the reconstructed optical flow and the reconstructed weight map. The decoding device can generate a reconstructed input frame based on the prediction frame.

Multi-scale optical flow for learned video compression

A video encoder may determine a set of quantization offset parameters for a group of scaled transform coefficients for a block of video data based on side information associated with the block of video data. The video encoder may further quantize the group of scaled transform coefficients for the block of video data to generate quantized transform coefficients for the block of video data based at least in part on the set of quantization offset parameters. The video encoder may further generate an encoded video bitstream based at least in part on the quantized transform coefficients for the block of video data.

Learned low-complexity adaptive quantization for video compression

Techniques are described herein for coding video content using recurrent-based machine learning tools. A device can include a neural network system including encoder and decoder portions. The encoder portion can generate output data for the current time step of operation of the neural network system based on an input video frame for a current time step of operation of the neural network system, reconstructed motion estimation data from a previous time step of operation, reconstructed residual data from the previous time step of operation, and recurrent state data from at least one recurrent layer of a decoder portion of the neural network system from the previous time step of operation. A decoder portion of the neural network system can generate, based on the output data and recurrent state data from the previous time step of operation, a reconstructed video frame for the current time step of operation.

Video compression using recurrent-based machine learning systems

Techniques are described for learned bidirectional predicted frame (B-frame) coding. An example method can include receiving a residual associated with a frame of a current time step; determining first motion information for a first reference frame associated with a first time step and second motion information for a second reference frame associated with a second time step, wherein the current time step is after the first time step and before the second time step; determining third motion information for the frame based on the first motion information and second motion information; generating a predicted frame based on the third motion information, first reference frame and second reference frame; and generating, using the predicted frame and residual, a reconstructed B-frame for the current time step, the reconstructed B-frame representing the frame.

Growing community of inventors

San Diego, CA, United States of America

Reza Pourreza