Afterimage Detection Using Integrated Frame Gradation Data

The patent merges the functionality of multiple separate memories into a single memory structure. The first memory stores previous frame data and the second memory stores similarity information, but both are accessed and processed within a unified memory management framework, reducing the overall number of memory components needed while maintaining detection accuracy

The patent makes the memory system multi-functional by having it perform both frame storage and similarity calculation functions. The same memory resources are used to store gradation data from previous frames and to compute similarity metrics, eliminating the need for dedicated separate memory blocks for each function

The patent performs preliminary actions by pre-calculating and storing similarity information between frames in the second memory before the actual afterimage detection is needed. This pre-computation of similarity metrics allows the detection algorithm to quickly identify candidate regions without performing comprehensive real-time comparisons, thus maintaining high detection accuracy while improving processing speed

The patent segments the frame comparison process into distinct stages: first storing raw frame data, then separately computing similarity information, and finally using both to detect afterimage regions. This segmentation allows each stage to be optimized independently, with similarity calculations performed in advance rather than during the critical detection phase