Disney Enterprises, Inc.

Eth Zurich

Jared McPhillen

Mark Arana

Kevin Liao

Christopher Richard Schroers

Scott C Labrozzi

Erika Varis Doggett

Abdelaziz Djelouah

Yuanyi Xue

Romann Matthew Weber

Joaquim Campos

Mauro Bamert

Stephan Marcel Mandt

Leonhard Markus Helminger

Simon Meierhans

According to one implementation, an image compression system includes a computing platform having a hardware processor and a system memory storing a software code. The hardware processor executes the software code to receive an input image, transform the input image to a latent space representation of the input image, and quantize the latent space representation of the input image to produce multiple quantized latents. The hardware processor further executes the software code to encode the quantized latents using a probability density function of the latent space representation of the input image, to generate a bitstream, and convert the bitstream into an output image corresponding to the input image. The probability density function of the latent space representation of the input image is obtained based on a normalizing flow mapping of one of the input image or the latent space representation of the input image.

Image compression using normalizing flows

Systems and methods for predicting a target set of pixels are disclosed. In one embodiment, a method may include obtaining target content. The target content may include a target set of pixels to be predicted. The method may also include convolving the target set of pixels to generate an estimated set of pixels. The method may include matching a second set of pixels in the target content to the target set of pixels. The second set of pixels may be within a distance from the target set of pixels. The method may include refining the estimated set of pixels to generate a refined set of pixels using a second set of pixels in the target content.

Systems and methods for image compression at multiple, different bitrates

A method, comprises monitoring a encoding process of a source video file performed by an encoder; obtaining an encoding decision parameter used to encode a picture of the source video file during the encoding process; comparing the encoding decision parameter to a threshold; based on the step of comparing, identifying the picture as a candidate picture for a visual defect or coding error; and storing a timestamp of the candidate picture.

Heuristic detection of potential digital media artifacts and defects in digital media assets

A data processing system includes a computing platform having a hardware processor and a memory storing a data compression software code. The hardware processor executes the data compression software code to receive a series of compression input data and encode a first compression input data of the series to a latent space representation of the first compression input data. The data compression software code further decodes the latent space representation to produce an input space representation of the first compression input data corresponding to the latent space representation, and generates f refined latent values for re-encoding the first compression input data based on a comparison of the first compression input data with its input space representation. The data compression software code then re-encodes the first compression input data using the refined latent values to produce a first compressed data corresponding to the first compression input data.

Content adaptive optimization for neural data compression

Systems and methods are disclosed for generating a latent space residual. A computer-implemented method may use a computer system that includes non-transient electronic storage, a graphical user interface, and one or more physical computer processors. The computer-implemented method may include: obtaining a target frame, obtaining a reconstructed frame, encoding the target frame into a latent space to generate a latent space target frame, encoding the reconstructed frame into the latent space to generate a latent space reconstructed frame, and generating a latent space residual based on the latent space target frame and the latent space reconstructed frame.

Systems and methods for generating a latent space residual

Systems and methods are disclosed for reconstructing a frame. A computer-implemented method may use a computer system that includes non-transient electronic storage, a graphical user interface, and one or more physical computer processors. The computer-implemented method may include: obtaining one or more reference frames from non-transient electronic storage, generating one or more displacement maps based on the one or more reference frames and a target frame with the physical computer processor, generating one or more warped frames based on the one or more reference frames and the one or more displacement maps with the physical computer processor, obtaining a conditioned reconstruction model from the non-transient electronic storage, and generating one or more blending coefficients and one or more reconstructed displacement maps by applying the one or more displacement maps, the one or more warped frames, and a target frame to the conditioned reconstruction model with the physical computer processor.

Systems and methods for reconstructing frames

Systems and methods for distortion removal at multiple quality levels are disclosed. In one embodiment, a method may include receiving training content. The training content may include original content, reconstructed content, and training distortion quality levels corresponding to the reconstructed content. The reconstructed content may be derived from distorted original content. The method may also include training distortion quality levels corresponding to the reconstructed content. The method may further include receiving an initial distortion removal model. The method may include generating a conditioned distortion removal model by training the initial distortion removal model using the training content. The method may further include storing the conditioned distortion removal model.

Systems and methods for distortion removal at multiple quality levels

Systems and methods are provided for manually or automatically identifying delineating scenes in media content. Scene categorization parameters setting forth thematic and/or cinematic criteria are used to analyze frames or groups of pictures (GOPs) making up the media content. Those frames or GOPs having one or more characteristics that meet the thematic and/or cinematic criteria may be deemed to belong or be a part of a particular scene in the media content. Downstream applications or processes can be optimized based on the scene segmentation.

Systems and methods for scene categorization

Systems and methods are provided for transcoding media content in accordance with thematic and/or cinematic characteristics of scenes. In particular, compression techniques may be selected to optimize compression of scenes based on their thematic and/cinematic characteristics. Scene boundary identifiers, such as scene in and out points may be identified, and the corresponding thematic and/or cinematic characteristic meta-data can be sent to a transcoder for transcoding. In this way, the specific characteristics of each scene in a movie can be accounted for during transcoding. This can result in transcoded versions of the movie that have better audio/visual qualities, and that can be encoded at a lower overall bit rate.

Systems and methods for profile based media segment rendering

Systems and methods are provided for re-transcoding portions of a digital media file while maintaining key-frame alignment across a set of digital media files having the same content. Instead of re-transcoding the entire set of digital media files if a portion of a single video fails quality, the disclosed systems and methods allow only the failing portion of the video to be re-transcoded.

Efficient re-transcoding of key-frame-aligned unencrypted assets

Systems and methods are provided for transcoding a set of key frame aligned output files from a source file. A key frame list file is generated while transcoding a first output file of the set. Subsequent output files of the set may be created non-sequentially and in parallel using the key frame list file.

Key frame aligned transcoding using key frame list file

Systems and methods for delivery of localized media assets to a content playback device are disclosed. A file server stores a base file and localized file associated with a localized media content. The base file and localized file each include one or more fragments. A web server delivers base file fragments and localized file fragments from the file server to the content playback device. Playback logic contained in one or more playlists determines an order of streaming each of the plurality of base file fragments and each of the plurality of localized file fragments for playback of the localized media content.

Systems and methods for delivery of localized media assets

Systems and methods are provided for reusing parts of previously created media assets when creating new media assets. Files created for previous media assets are evaluated when creating a manifest file for a new media asset. If a needed file was previously created, the manifest file for the new media asset references the preexisting file. This avoids the need to recreate files when creating new media assets.

Efficient re-use of a media asset

Network condition prediction and multimedia streaming consumption prediction are provided. The prediction may be based on a device's prior location, behavior, and statistics thereof. By gathering location data from users anonymously and securely, a virtual location network with millions of nodes are provided. Each virtual location, at a given time, is stored with associated network metrics gathered from various devices in a database. The database may comprise a probabilistic model and a behavioral model tracking device metrics.

Growing community of inventors

Glendale, CA, United States of America

Jared McPhillen