Google Inc.

University of Houston System

George Dan Toderici

Sanketh Shetty

Nicholas M Johnston

Weilong Yang

Sudheendra Vijayanarasimhan

Balakrishnan Varadarajan

Apostol Ivanov Natsev

David Charles Minnen

Sung Jin Hwang

Nitin Khandelwal

Saurabh Singh

Michele Mae Covell

Sami Ahmad Abu-El-Haija

Johannes Balle

Joel Eric Shor

Eirikur Thor Agustsson

Damien Vincent

Jay Yagnik

Hrishikesh Balkrishna Aradhye

Troy Chinen

Tomas Izo

Nicola Muscettola

Min-Hsuan Tsai

Phuong B Le

WeiHsin Gu

Fabian Julius Mentzer

Oriol Vinyals

Jyrki Antero Alakuijala

Omid Madani

Rahul Sukthankar

Luciano Sbaiz

Rajat Monga

Susanna Maria Ricco

Sean O'Malley

Abhinav Shrivastava

John Burge

Yue Hei Ng

Matthew John Hausknecht

Michael Tobias Tschannen

Philip Andrew Chou

Alexandru Marius Pasca

Ioannis A Kakadiaris

Mehmet Emre Sargin

Alex Sukhanov

Andrej Karpathy

Jasper Snoek

Theoharis Theoharis

Georgios Passalis

Berivan Isik

Susana Ricco

Johannes Ballé

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for compressing video data. In one aspect, a method comprises: receiving a video sequence of frames; generating, using a flow prediction network, an optical flow between two sequential frames, wherein the two sequential frames comprise a first frame and a second frame that is subsequent the first frame; generating from the optical flow, using a first autoencoder neural network: a predicted optical flow between the first frame and the second frame; and warping a reconstruction of the first frame according to the predicted optical flow and subsequently applying a blurring operation to obtain an initial predicted reconstruction of the second frame.

Video compression using optical flow

An example computer-implemented method for determining a perceptual quality of a subject video content item is provided. The example method can include inputting a subject frame set from the subject video content item into a first machine-learned model. The example method can also include generating, using the first machine-learned model, a feature based at least in part on the subject frame set. The example method can also include outputting, using a second machine-learned model, a score indicating the perceptual quality of the subject video content item based at least in part on the feature.

Method for constructing a perceptual metric for judging video quality

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an encoder neural network configured to receive a data item and to process the data item to output a compressed representation of the data item. In one aspect, a method includes, for each training data item: processing the data item using the encoder neural network to generate a latent representation of the training data item; processing the latent representation using a hyper-encoder neural network to determine a conditional entropy model; generating a compressed representation of the training data item; processing the compressed representation using a decoder neural network to generate a reconstruction of the training data item; processing the reconstruction of the training data item using a discriminator neural network to generate a discriminator network output; evaluating a first loss function; and determining an update to the current values of the encoder network parameters.

High-fidelity generative image compression

A system and methodology provide for annotating videos with entities and associated probabilities of existence of the entities within video frames. A computer-implemented method identifies an entity from a plurality of entities identifying characteristics of video items. The computer-implemented method selects a set of features correlated with the entity based on a value of a feature of a plurality of features, determines a classifier for the entity using the set of features, and determines an aggregation calibration function for the entity based on the set of features. The computer-implemented method selects a video frame from a video item, where the video frame having associated features, and determines a probability of existence of the entity based on the associated features using the classifier and the aggregation calibration function.

Feature-based video annotation

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving, by a neural network (NN), a dataset for generating features from the dataset. A first set of features is computed from the dataset using at least a feature layer of the NN. The first set of features i) is characterized by a measure of informativeness; and ii) is computed such that a size of the first set of features is compressible into a second set of features that is smaller in size than the first set of features and that has a same measure of informativeness as the measure of informativeness of the first set of features. The second set of features if generated from the first set of features using a compression method that compresses the first set of features to generate the second set of features.

Learning compressible features

A computer-implemented method for selecting representative frames for videos is provided. The method includes receiving a video and identifying a set of features for each of the frames of the video. The features including frame-based features and semantic features. The semantic features identifying likelihoods of semantic concepts being present as content in the frames of the video. A set of video segments for the video is subsequently generated. Each video segment includes a chronological subset of frames from the video and each frame is associated with at least one of the semantic features. The method generates a score for each frame of the subset of frames for each video segment based at least on the semantic features, and selecting a representative frame for each video segment based on the scores of the frames in the video segment. The representative frame represents and summarizes the video segment.

Selecting and presenting representative frames for video previews

Example embodiments of the present disclosure relate to systems and methods for compressing attributes of volumetric and hypervolumetric datasets. An example system performs operations including obtaining a reference dataset comprising attributes indexed by a domain of multidimensional coordinates; subdividing the domain into a plurality of blocks respectively associated with a plurality of attribute subsets; inputting, to a local nonlinear operator, a latent representation for an attribute subset associated with at least one block of the plurality of blocks; obtaining, using the local nonlinear operator and based on the latent representation, an attribute representation of one or more attributes of the attribute subset; and updating the latent representation based on a comparison of the attribute representation and the reference dataset.

Learned volumetric attribute compression using coordinate-based networks

Artificial image generation may include obtaining a source image, identifying quantization information from the source image, wherein identifying the quantization information includes identifying multiresolution quantization interval information from the source image, generating a restoration filtered image by restoration filtering the source image, generating a constrained restoration filtered image by constraining the restoration filtered image based on the quantization information, obtaining an unconstrained artificial image based on the constrained restoration filtered image and a generative artificial neural network obtained using a generative adversarial network, obtaining the artificial image by constraining the unconstrained artificial image based on the quantization information, and outputting the artificial image.

Quantization constrained neural image coding

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for compressing and decompressing data. In one aspect, a method comprises: processing data using an encoder neural network to generate a latent representation of the data; processing the latent representation of the data using a hyper-encoder neural network to generate a latent representation of an entropy model; generating an entropy encoded representation of the latent representation of the entropy model; generating an entropy encoded representation of the latent representation of the data using the latent representation of the entropy model; and determining a compressed representation of the data from the entropy encoded representations of: (i) the latent representation of the data and (ii) the latent representation of the entropy model used to entropy encode the latent representation of the data.

Data compression using conditional entropy models

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image compression and reconstruction. A request to generate an encoded representation of an input image is received. The encoded representation of the input image is then generated. The encoded representation includes a respective set of binary codes at each iteration. Generating the set of binary codes for the iteration from an initial set of binary includes: for any tiles that have already been masked off during any previous iteration, masking off the tile. For any tiles that have not yet been masked off during any of the previous iterations, a determination is made as to whether a reconstruction error of the tile when reconstructed from binary codes at the previous iterations satisfies an error threshold. When the reconstruction quality satisfies the error threshold, the tile is masked off.

Stop code tolerant image compression neural networks

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image compression and reconstruction. An image encoder system receives a request to generate an encoded representation of an input image that has been partitioned into a plurality of tiles and generates the encoded representation of the input image. To generate the encoded representation, the system processes a context for each tile using a spatial context prediction neural network that has been trained to process context for an input tile and generate an output tile that is a prediction of the input tile. The system determines a residual image between the particular tile and the output tile generated by the spatial context prediction neural network by process the context for the particular tile and generates a set of binary codes for the particular tile by encoding the residual image using an encoder neural network.

Tiled image compression using neural networks

Growing community of inventors

Mountain View, CA, United States of America

George Dan Toderici