Video coding method and apparatus using block vectors with adaptive spatial resolution
Adaptive signaling of block vector spatial resolution and differences addresses the inefficiencies in conventional video encoding for screen content, particularly in super high-definition scenarios, by enabling sub-pixel unit representation and improving encoding efficiency.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- HYUNDAI MOTOR CO LTD
- Filing Date
- 2022-03-28
- Publication Date
- 2026-05-22
AI Technical Summary
Conventional video encoding methods for screen content, such as HEVC and VVC, fail to efficiently encode super high-definition content due to the lack of consideration for sub-pixel units in block vector representation, leading to reduced encoding efficiency.
A method for adaptively signaling the spatial resolution of block vectors and the sign of block vector differences, allowing for sub-pixel unit representation, by generating and combining block vector predictors and differences to improve encoding efficiency.
Improves coding efficiency by adaptively signaling the spatial resolution of block vectors, enhancing encoding performance for super high-definition screen content.
Smart Images

Figure 0007863794000006 
Figure 0007863794000007 
Figure 0007863794000008