Post-filtering of a video using multiple sample arrays
By incorporating auxiliary sample arrays from the same time frame, the neural network post-filtering process improves video quality and reduces computational load and signaling overhead, addressing the limitations of existing NNPF technologies.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
- Filing Date
- 2026-01-02
- Publication Date
- 2026-07-09
AI Technical Summary
Existing neural network post-filtering (NNPF) technologies for video processing face challenges in achieving a balance between high-quality picture generation, low computational load, and low signaling overhead, particularly when handling multiple layers of video data.
The use of auxiliary sample arrays from the same time frame, such as different resolution, view position, or depth maps, to enhance the filtering process in neural network post-filtering, allowing for improved quality and computational efficiency by selectively filtering specific image areas or enhancing objects based on object masks.
This approach enhances the quality of filtered video output while reducing computational load and signaling overhead, providing a better trade-off in video post-filtering processes.
Smart Images
Figure EP2026050028_09072026_PF_FP_ABST