Methods and devices for neural image compression to replace computer-readable terminals and recording media.
VN10062255BActive Publication Date: 2026-11-02TENCENT AMERICA LLC
View PDF 0 Cites 0 Cited by
Patent Information
- Application Number
- VN1202207211
- Authority / Receiving Office
- VN · VN
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-10-14
- Publication Date
- 2026-11-02
- Estimated Expiration
- 2041-10-14
Abstract
The invention refers to an end-to-end (E2E) replacement image compression (NIC) method based on neural networks implemented by at least one processor and includes receiving an input image for an E2E NIC software frame, determining the order size of the input image to indicate the learning rate of the training model, determining a replacement image based on the training model, encoding the replacement image in place of the input image to generate a bitstream, and mapping the replacement image to the bitstream to create a compressed representation. Additionally, the order size can be determined by a scheduler and changed during training of the training model. The image can also be split into sheets, with a scheduler assigned to each sheet, and each sheet encoded in place of the entire input image.
Need to check novelty before this filing date? Find Prior Art
