Image Block Processing in DDR to Reduce Transmission Delay

The patent divides the entire frame of video source image data into multiple image blocks (e.g., 16x16 pixels) that can be processed independently. The DDR is organized with each row containing data for multiple image blocks, allowing the system to read and process partial frames (individual blocks or groups of blocks) rather than waiting for complete frame accumulation. This segmentation enables incremental compression of image blocks as they become available in the DDR, significantly reducing transmission delay while maintaining compression quality.

The patent implements preliminary organization of the DDR memory structure where each row is configured to store source image data for a specific number of image blocks. The system pre-calculates and sets the number of rows to be read based on the required number of image blocks for compression. This preliminary configuration allows the JPEG compression module to immediately begin processing as soon as the required data is written, without waiting for the entire frame to be complete, thereby improving processing efficiency while ensuring data completeness for each image block.

The patent performs partitioning of the image frame at the memory storage level rather than at the compression processing level. By organizing the DDR to naturally contain segmented image block data and configuring the read operation to retrieve exactly the required number of rows corresponding to the needed image blocks, the system avoids the overhead of dynamic partitioning logic in the JPEG compression module. This approach reduces transmission delay by enabling selective reading of partial frames while minimizing system resource waste by eliminating redundant partitioning operations.