Person Detection via Occupancy Rate Analysis in Moving Body Frames
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current monitoring systems face challenges in executing accurate person detection from images due to high computational requirements and frequent erroneous detections of non-human objects as persons, leading to increased server load and processing time for criminal detection.
Innovation Solution
An image processing apparatus and method that performs a pre-determination process to set a person detection frame around a moving body region and a final determination process to analyze the distribution state within the frame, using a lightweight calculation to accurately identify persons and reduce erroneous detections by calculating occupancy rates and feature quantities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If person detection process is executed in each monitoring camera, then detection accuracy is improved, but computational load and processing time increase
Solution Approach 1:
The patent divides the monitoring system into two segments: monitoring cameras that perform only moving body detection, and information terminals that perform person detection. This segmentation allows each device to have optimized processing capabilities, reducing the computational load on cameras while maintaining accurate person detection at the information terminal level.
Solution Approach 2:
The patent introduces an intermediary component (information terminal) that receives moving body detection results from cameras and performs the more computationally intensive person detection process. This intermediary handles the complex analysis, allowing cameras to remain simple devices with lower computational requirements.
2Measurement precision
If face detection is used for person detection, then detection accuracy is improved, but images without faces cannot be detected
Solution Approach 1:
The patent performs moving body detection as a preliminary action before person detection. By first identifying moving bodies and then performing person detection on those specific regions, the system can detect persons even when faces are not visible, while still maintaining accuracy through the subsequent person detection process that analyzes body features and patterns.
Data Source
AI summary
A pre-determination process is executed for setting a person determination detection frame in a moving body detection frame surrounding a moving body region detected from an image and determining whether the image in the person determination detection frame is a person and a final determination process of analyzing a distribution state in a moving body detection frame determined to include a person in the pre-determination process and finally determining whether the image in the person determination detection frame is a person based on an analysis result. In the final determination process, an occupancy rate in a specific region in the moving body detection frame determined to include a person in the pre-determination process is calculated, and the image in the person determination detection frame is determined to be a person based on the occupancy rate.