US 11,638,007 B2 · Codebook generation for cloud-based video applications

Abstract

Techniques are disclosed for the improvement of vector quantization (VQ) codebook generation. The improved codebooks may be used for compression in cloud-based video applications. VQ achieves compression by vectorizing input video streams, matching those vectors to codebook vector entries, and replacing them with indexes of the matched codebook vectors along with residual vectors to represent the difference between the input stream vector and the codebook vector. The combination of index and residual is generally smaller than the input stream vector which they collectively encode, thus providing compression. The improved codebook may be generated from training video streams by grouping together similar types of data (e.g., image data, motion data, control data) from the video stream to generate longer vectors having higher dimensions and greater structure. This improves the ability of VQ to remove redundancy and thus increase compression efficiency. Storage space is thus reduced and video transmission may be faster.

Claims20

Dependent claims

A method for generating a codebook for vector quantization of digital video content, the method comprising:

performing entropy decoding on a pre-compressed training video stream to provide a decoded video stream;
partitioning the decoded video stream into a plurality of data segment clusters including a first data segment cluster comprising a plurality of image data segments, a second data segment cluster comprising a plurality of motion data segments, and a third data segment cluster comprising a plurality of control data segments;
grouping at least one of the plurality of data segment clusters to form a data vector that is longer than each data segment in the grouped at least one data segment cluster;
generating a codebook entry that is at least partially based on the data vector;
and
associating an index value with the generated codebook entry, wherein the index value is shorter than, and provides a compressed representation of, the generated codebook entry.

3 more elements

A computer program product comprising one or more non-transitory computer readable media encoded with instructions that, when executed by one or more processors, cause a process for generating a codebook for vector quantization of digital video content to be invoked, the process comprising:

performing entropy decoding on a pre-compressed training video stream to provide a decoded video stream;
partitioning the decoded video stream into a plurality of data segment clusters including a first data segment cluster comprising a plurality of image data segments, a second data segment cluster comprising a plurality of motion data segments, and a third data segment cluster comprising a plurality of control data segments;
grouping a particular one of the plurality of data segment clusters to form a data vector that is longer than each data segment in the grouped particular data segment cluster;
generating a codebook entry that is at least partially based on the data vector;
and
associating an index value with the generated codebook entry, wherein the index value is shorter than, and provides a compressed representation of, the generated codebook entry.

3 more elements

A system for generating a codebook for vector quantization of digital video content, the system comprising a memory and one or more processors configured to:

perform entropy decoding on a pre-compressed video stream to provide a decoded video stream;
partition the decoded video stream into a plurality of data segment clusters including a first data segment cluster comprising a plurality of image data segments, a second data segment cluster comprising a plurality of motion data segments, and a third data segment cluster comprising a plurality of control data segments;
group at least one of the plurality of data segment clusters to form a data vector that is longer than each data segment in the grouped at least one data segment cluster;
generate a codebook entry that is at least partially based on the data vector;
and
associate an index value with the generated codebook entry, wherein the index value is shorter than, and provides a compressed representation of, the generated codebook entry.

3 more elements

Dependencies are inferred from the claim text.

Codebook generation for cloud-based video applications

Life of this patent

Codebook generation for cloud-based video applications

Life of this patent

Abstract

Claims20