Back to results
US 12,641,301 B2Granted
Context-aware error concealment to improve inference accuracy
AssigneeNVIDIA Corporation
InventorsSwapnil Jagdish Rathi, Bhushan Rupde, Kaustubh Purandare
Life of this patent
Estimating the term…
Abstract
In various examples, systems and methods are disclosed relating to context-aware error concealment to improve inference accuracy are disclosed. A system can identify a frame of a video stream and determine that the frame comprises corrupted or lost data. The system can generate a corrected frame by applying an error concealment function selected based at least on a location of the corrupted or lost data in the frame and a region of interest in the frame.
Claims15
One or more processors comprising:
- one or more circuits to:
- identify a frame of a video stream, the frame having at least one region of interest;
- determine that the frame comprises corrupted or lost data;
- select a first error concealment function for a first portion of the corrupted or lost data within the at least one region of interest, and a second error concealment function for a second portion of the corrupted or lost data outside of the at least one region of interest in the frame;
- and
- generate a corrected frame by applying the first error concealment function to the first portion of the corrupted or lost data and by applying the second error concealment function to the second portion of the corrupted or lost data.
3 more elements
A system, comprising:
- one or more processors to:
- receive a request to process a video stream comprising a plurality of frames;
- apply a first error concealment function and a second error concealment function to at least one frame of the plurality of frames, wherein the first error concealment function is selected for a first portion of corrupted or lost data of the at least one frame within at least one region of interest, and the second error concealment function is selected for a second portion of the corrupted or lost data of the at least one frame outside of the at least one region of interest;
- and
- execute a machine-learning model using the at least one frame as input for the request.
2 more elements
A method, comprising:
- identifying, using one or more processors, a frame of a video stream, the frame having at least one region of interest;
- determining, using the one or more processors, that the frame comprises corrupted or lost data;
- selecting, using the one or more processors, a first error concealment function for a first portion of the corrupted or lost data within the at least one region of interest, and a second error concealment function for a second portion of the corrupted or lost data outside of the at least one region of interest in the frame;
- and
- generating, using the one or more processors, a corrected frame by applying the first error concealment function to the first portion of the corrupted or lost data and the second error concealment function to the second portion of the corrupted or lost data.
2 more elements
Dependencies are inferred from the claim text.