Patent · US Active

K-nearest neighbor model-based content adaptive encoding parameters determination

US10721475B2 · kind B2 · utility

5Cited by
2References
26Claims
0Family size

Assignee

Inventors

Key dates

Filing dateAug 31, 2018
Grant dateJul 21, 2020
Priority date
Expiry dateSep 15, 2038

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06N20/20
  • WIPO fieldAudio-visual technology
  • WIPO sectorElectrical engineering

Abstract

In one example, a segment of media content, a predefined maximum bitrate, and a predefined maximum quality measure may be received. K-NN having similar bitrate and quality data to the segment of media content may be determined using a trained k-nearest neighbor model. The K-NN are determined from a training set of media content segments. Further, at least one content adaptive encoding parameter for the segment of media content may be determined within the predefined maximum bitrate and the predefined maximum quality measure using the determined k-nearest neighbor. The at least one content adaptive encoding parameter is inputted to an encoder to encode the segment of media content.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.