Image-Based Body Angle Detection for Abduction Risk Alerts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a risk of individuals being subjected to unintended actions, such as kidnapping or abduction, which existing image processing systems fail to effectively detect.
Innovation Solution
An image processing apparatus that analyzes multiple images of an individual taken at different times to detect changes in body angle and the presence of an attention target, such as another person or object, and outputs information when specific criteria are met, indicating a potential risk.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing image processing systems are used, then general image analysis is performed, but detection of unintended actions such as kidnapping or abduction is not achieved
Solution Approach 1:
The patent applies local quality by focusing image processing on specific body parts (head, shoulders, hands) rather than analyzing the entire body uniformly. This localized approach enables detection of specific risk indicators such as hand placement on the head or abnormal body angles, thereby improving detection accuracy for unintended actions while maintaining system adaptability
Solution Approach 2:
The patent utilizes parameter changes by monitoring temporal variations in body angle and hand position parameters across multiple images. By detecting changes in these specific parameters over time (e.g., head angle change exceeding a threshold, hand moving to cover the head), the system can identify unintended actions like kidnapping or abduction, resolving the contradiction between reliable detection and scenario-specific adaptability
2Measurement precision
If multiple images are processed to detect body angle changes, then detection accuracy for unintended actions is improved, but processing time and computational load increase
Solution Approach 1:
The patent extracts only the essential features (body angle, hand position) from multiple images rather than processing complete image data. By focusing on these specific extracted parameters and comparing their temporal changes, the system achieves accurate detection of unintended actions while significantly reducing processing time and computational requirements
Solution Approach 2:
The patent applies partial action by processing only the necessary subset of image data (specific body parts and angles) rather than analyzing entire images in full detail. This selective processing approach maintains measurement precision for critical parameters while minimizing the time loss associated with processing all image data
Data Source
AI summary
An image processing apparatus includes an image processing unit and a determination unit. The image processing unit acquires a plurality of images (e.g., frame images) generated by an image capturing apparatus. Then, the image processing unit processes an image including the same person (a first person) among the plurality of images, and thereby detects a change in an angle of a body of the first person. Moreover, the image processing unit detects, at least when magnitude of the change satisfies a criterion, information relating to a second person or an object (an attention target) existing in a direction in which the body leans. The determination unit outputs predetermined information by use of a processing result of the image processing unit. As one example, the determination unit outputs predetermined information, when a first criterion relating to at least one of the attention target and the first person described above is satisfied.