Parallel Video Processing Without Frame Memory

The patent segments the video signal into multiple frames and processes them in parallel using multiple processors. Each processor handles specific frames independently, eliminating the need for a single large frame memory. The frame memory is divided into multiple smaller memory units, one for each processor, reducing overall device complexity while maintaining parallel processing capability.

The patent transitions from a sequential processing model to a parallel processing model by introducing a temporal dimension. Instead of processing frames sequentially in a single processor, the system distributes frames across multiple processors simultaneously. This dimensional change allows parallel processing without requiring a single large frame memory, as each processor only needs access to its assigned frames.

The patent segments the frame memory into multiple smaller memory units, with each processor having its own dedicated memory for storing assigned frames. This segmentation approach eliminates the need for a single large expensive frame memory. Multiple smaller memory units are less costly than one large memory, while still providing sufficient storage for parallel processing of multiple frames simultaneously.

The patent uses smaller, less expensive frame memories that are sufficient for individual frame storage rather than requiring a large, expensive dedicated frame memory. Each processor has its own small memory unit that can be independently optimized for cost-effectiveness. This approach trades the single large memory requirement for multiple smaller memory units, reducing overall system cost while maintaining the necessary processing capability.