Image Processing Apparatus for Adaptive Face Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems for face detection in surveillance cameras face high computational loads, leading to missed detections when images of varying sizes are processed, especially for objects that may disappear from the screen due to large motion, as they require different resolution calculations.
Innovation Solution
An image processing apparatus that sequentially reduces images based on a predetermined reduction ratio and frame rate, allowing for efficient face detection by adapting the resolution and processing rate according to the object's size and distance from the camera, prioritizing detection of larger objects displayed on the screen.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If face detection is performed on images of various resolutions to detect objects of various sizes, then detection coverage is improved, but calculation amount increases
Solution Approach 1:
The patent divides the detection process into multiple stages with different resolution levels. Images are processed at multiple resolutions (e.g., original resolution and reduced resolution), with each resolution level handling different size ranges of detection targets. This segmentation allows the system to detect various sized objects while reducing total computational load by not processing all images at full resolution.
Solution Approach 2:
The patent applies different processing qualities to different regions or resolution levels. Full-resolution processing is applied only where necessary (for larger objects or critical detection zones), while reduced-resolution processing is applied to other areas. This local quality approach optimizes the balance between detection accuracy and computational resources.
2Measurement precision
If face detection is performed on large-size images, then detection accuracy for large objects is improved, but processing time increases
Solution Approach 1:
The patent dynamically adjusts the resolution level used for processing based on the detection needs. Rather than using a fixed resolution for all cases, the system adaptively selects processing resolutions based on factors such as object size, detection confidence requirements, and computational resource availability. This dynamic approach allows accurate detection when needed while reducing processing time when full accuracy is not critical.
3Reliability
If detection priority is given to objects displayed large (near camera), then important objects are detected timely, but computational resources are concentrated on specific regions
Solution Approach 1:
The patent performs preliminary detection at reduced resolution to identify regions of interest before conducting detailed detection at full resolution. This preliminary action allows the system to quickly identify potential targets and allocate computational resources efficiently, ensuring that objects displayed large (which are more likely to be important) receive timely attention while maintaining the ability to detect other objects at appropriate resolution levels.
Data Source
AI summary
An image is input on a frame unit basis, the input image is sequentially reduced, and an object is detected from the input image and the reduced image at a frame rate according to a reduction ratio of the reduced image to the input image, thereby decreasing an amount of calculations necessary to detect the object from the image.