Selective Video Block Processing to Reduce AI Bandwidth Use
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional AI models process redundant data between successive video frames, leading to inefficient resource utilization and high bandwidth consumption due to repetitive computations of identical or minimally changed image data.
Innovation Solution
The method involves dividing images into blocks and processing only those blocks that satisfy a predefined difference threshold, splicing the results of these blocks with the processed image blocks to generate the final output, thereby reducing unnecessary computations and bandwidth usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the AI model processes the entire image frame for each video frame, then the processing completeness is maintained, but the bandwidth consumption and computational load increase significantly due to redundant data processing
Solution Approach 1:
The image frame is divided into multiple blocks, and the AI model processes only the changed blocks rather than the entire frame. This segmentation approach maintains processing completeness for changed regions while eliminating redundant computation on unchanged regions, thereby reducing bandwidth consumption and computational load.
Solution Approach 2:
The patent extracts only the changed blocks from the current frame by comparing with the previous frame, and feeds only these extracted changed blocks to the AI model. This extraction principle eliminates unnecessary redundant data transmission and processing, reducing bandwidth consumption while maintaining processing effectiveness.
2Reliability
If the AI model repeatedly processes the entire image frame for each video frame, then all image data is processed, but the resource utilization efficiency decreases due to redundant computations
Solution Approach 1:
By segmenting the frame into blocks and identifying only changed blocks through comparison, the system processes only necessary regions. This ensures complete processing of changed data while improving resource utilization efficiency by eliminating redundant computations on unchanged blocks.
Solution Approach 2:
The patent applies partial action by processing only the changed blocks rather than the entire frame. This partial processing approach maintains adequacy for the actual information needs while significantly improving resource utilization efficiency by avoiding excessive computation on redundant data.
3Loss of information
If the entire image frame is transmitted and processed for each video frame, then no data is lost, but the bandwidth requirements and computational resources increase substantially
Solution Approach 1:
The patent extracts only the changed blocks from the current frame by comparing with the previous frame, and transmits only these extracted changed blocks for AI processing. This extraction approach ensures no information loss regarding changes while substantially reducing data transmission volume and bandwidth requirements.
Solution Approach 2:
By segmenting the frame into blocks and transmitting only the changed segments, the patent maintains data completeness for changed regions while reducing overall data transmission volume. The segmentation enables selective transmission of only necessary information.
Data Source
AI summary
A data processing method includes inputting first blocks from a plurality of blocks of a divided to-be-processed image into an image processing model, to obtain block processing results of the first blocks, and performing splicing processing on the block processing results of the first blocks and block processing results of second blocks of the processed image, to obtain an image processing result of the to-be-processed image. The first blocks are blocks of the to-be-processed image, which satisfy a processing condition. The processed image and the to-be-processed image are divided in a same division method. The second blocks are blocks of the to-be-processed image which do not satisfy the processing condition, compared to the processed image.