Image Encoding Apparatus for Multi-Region Code Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image signal encoding techniques for monitoring cameras face challenges in maintaining intended image quality when multiple specific areas exist in a frame, as they either exceed the target code amount or fail to provide sufficient image quality, especially when encoding human faces.
Innovation Solution
An image processing apparatus that extracts specific areas, evaluates their importance using a predetermined formula, and adjusts the encoding parameters to ensure the code amount for each frame does not exceed the target, by reducing the code amount for specific areas that exceed a threshold, thereby maintaining image quality without increasing the overall code amount.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If large code amounts are assigned to all specific areas to maintain image quality, then the image quality in specific areas is improved, but the code amount of the entire frame increases and exceeds the target value
Solution Approach 1:
The patent applies local quality by differentiating the treatment of specific areas based on their evaluation values. Instead of uniformly encoding all specific areas, the system selectively controls the degree of encoding for each specific area according to its evaluated importance, thereby maintaining high image quality for critical areas while managing overall code amount.
Solution Approach 2:
The patent changes the encoding parameter (degree of encoding) dynamically based on the evaluation value of each specific area. When the evaluation value exceeds a predetermined threshold, the degree of encoding is controlled to be a certain value or less, effectively adjusting the code amount allocation to balance image quality and code efficiency.
2Quantity of substance
If code amounts are assigned averagely to all specific areas to keep the total code amount within target, then the code amount control is simplified, but the specific area cannot obtain the intended image quality for identification
Solution Approach 1:
The patent implements local quality by evaluating each specific area individually and assigning different encoding priorities. The extraction unit identifies specific areas, the evaluation unit assigns evaluation values based on importance, and the degree of encoding is selectively controlled for areas exceeding the threshold, ensuring critical areas receive sufficient code allocation while maintaining overall code efficiency.
Solution Approach 2:
The patent employs feedback mechanisms where the evaluation unit continuously assesses the importance of specific areas, and this evaluation feeds back to control the encoding process. The determination of whether to limit the degree of encoding for a specific area is based on feedback from its evaluation value compared against a predetermined threshold, enabling dynamic adaptation to maintain both code efficiency and image quality.
3Productivity
If the area ratio of specific area to entire input image is low, then the overall compression ratio is improved, but when many specific areas exist, the intended image quality cannot be obtained for each area
Solution Approach 1:
The patent dynamically changes the encoding parameter (degree of encoding) based on the evaluation value of each specific area. When evaluation values exceed the threshold, the degree of encoding is constrained to maintain image quality, regardless of the number of specific areas or their area ratios. This allows the system to adapt to varying scene complexity while maintaining quality for important areas.
Solution Approach 2:
The patent applies local quality by treating each specific area according to its evaluated importance rather than uniformly. Even when many specific areas exist, the system identifies which areas require high quality representation and allocates encoding resources accordingly, ensuring that critical areas maintain intended image quality while allowing less critical areas to contribute to overall compression efficiency.
Data Source
AI summary
An area where a specific object is captured is extracted as a specific area from the image of a frame of interest, and the evaluation value of the specific area is obtained using a predetermined evaluation formula. It is determined whether the evaluation value of the specific area in a frame preceding the frame of interest has exceeded a predetermined threshold. When it is determined that the evaluation value of the specific area has exceeded the predetermined threshold, the frame of interest is encoded to set the code amount of the specific area in the image of the frame of interest to be smaller than that of the specific area in the image of the frame preceding the frame of interest.