Image Processing Scene Recognition Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video surveillance, frequent continuous scene transitions can lead to inaccurate recognition of scene categories, and target regions such as facial or license plate areas may be overexposed or underexposed due to poor lighting conditions, resulting in ineffective presentation.
Innovation Solution
A system and method that process images by obtaining current and historical images, determining confidence levels for scene categories, updating confidence thresholds based on target scene categories, and adjusting image processing techniques accordingly to ensure accurate scene recognition and optimal brightness of target regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If scene category recognition is performed on each image independently, then the recognition speed is fast, but the accuracy decreases due to frequent continuous scene transitions
Solution Approach 1:
The system performs preliminary scene category recognition on historical images before processing the current image. By pre-establishing the scene categories of previous frames, the system creates a temporal context that helps stabilize recognition accuracy during frequent scene transitions, while maintaining fast recognition speed through efficient use of pre-computed historical data
Solution Approach 2:
The system uses the recognized scene categories of historical images as feedback to adjust and improve the recognition of the current image. By comparing current recognition results with historical patterns and using this feedback loop, the system maintains high accuracy even during rapid scene changes without sacrificing recognition speed
2Manufacturing precision
If image processing is performed to improve target region brightness, then the presentation quality improves, but the processing complexity increases
Solution Approach 1:
The system applies different processing strategies to different regions of the image. Specifically, it identifies target regions (such as facial regions or license plate regions) and applies brightness adjustment only to these specific areas rather than processing the entire image uniformly. This localized approach improves target region presentation quality while minimizing overall processing complexity
Solution Approach 2:
The system adjusts brightness parameters selectively based on the identified scene category and target region characteristics. By changing only the necessary brightness parameters for specific regions rather than globally adjusting all image parameters, the system achieves improved presentation quality with reduced processing complexity
Data Source
AI summary
The present disclosure relates to a system for image processing. The system obtains an image, multiple historical images captured before the image, and multiple scene categories. For each scene category, the system generates a confidence level of the image belonging to the scene category. The system obtains an initial scene category of each historical image. The system determines a target scene category to which the multiple historical images belong based on initial scene categories of the multiple historical images. The system determines multiple updated confidence level thresholds by updating, based on the target scene category, at least a portion of multiple confidence level thresholds corresponding to the multiple scene categories. The system determines, based on the multiple updated confidence level thresholds and confidence levels of the image corresponding to the multiple scene categories, a final scene category of the image, and processes the image based on the final scene category.