Object-Aware Video Quality Control for Bandwidth and Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video transmission techniques struggle to reduce data amount effectively, as maintaining high image quality in gaze regions leads to increased data volume, while reducing overall image quality compromises recognition accuracy.
Innovation Solution
A video processing system with an image quality control apparatus that dynamically adjusts image quality based on object detection results, enhancing regions of interest and reducing quality in non-essential areas to manage data volume and maintain recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image quality of gaze region is always maintained high, then recognition accuracy is improved, but data amount increases
Solution Approach 1:
The video is divided into multiple regions, with the gaze region segmented from other regions. Different image quality levels are applied to different regions: high quality for the gaze region and low quality for non-gaze regions. This segmentation allows the system to maintain recognition accuracy in the important gaze region while reducing overall data transmission volume by degrading quality in less important areas.
Solution Approach 2:
Different image quality levels are applied locally to different regions of the video. The gaze region receives high image quality to ensure recognition accuracy, while other regions receive low image quality to reduce data amount. This local differentiation of quality resolves the contradiction by concentrating resources where they are most needed.
2Quantity of substance
If image quality of entire video is lowered, then data amount is reduced, but recognition rate decreases
Solution Approach 1:
Instead of uniformly lowering image quality across the entire video, the system applies local quality differentiation: the gaze region maintains high image quality to preserve recognition rate, while non-gaze regions are degraded to reduce data amount. This resolves the contradiction by protecting recognition performance in critical areas while achieving data reduction overall.
Solution Approach 2:
The video is segmented into gaze region and non-gaze region, allowing differential quality treatment. The gaze region is preserved at high quality to maintain recognition rate, while other segments are compressed to reduce data amount, thus resolving the contradiction between data reduction and recognition performance.
3Area of stationary object
If there are many gaze regions, then coverage is improved, but fewer regions can be degraded reducing data amount
Solution Approach 1:
The system dynamically adjusts which regions are designated as gaze regions based on detection results. When multiple gaze regions are detected, the system adaptively manages the quality allocation across these regions, potentially prioritizing certain regions or adjusting quality levels dynamically to balance coverage and data reduction goals.
Data Source
AI summary
A video processing system includes an image quality control apparatus and a detection apparatus. The image quality control apparatus includes an image quality control unit configured to control image quality of each region of a video, and a transmission unit configured to transmit, to the detection apparatus, the video of which the image quality is controlled. The detection apparatus includes a detection unit configured to detect information regarding an object in the video transmitted from transmission unit, and a notification unit configured to notify the image quality control apparatus of a detection result of the detection unit. The image quality control apparatus further includes a determination unit configured to determine the image quality of each region of the video according to the detection result notified from the notification unit, the image quality being controlled by the image quality control unit.