Human Intrusion Detection Using Multi-Part State Discrimination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video monitoring systems fail to differentiate between human bodies in vehicles and pedestrians, leading to false alarms when vehicles enter restricted areas.
Innovation Solution
A detection device that discriminates between different states of human bodies by detecting multiple portions using discriminators, such as face, head, upper body, and whole body, and determines whether the detected object is a pedestrian or a vehicle passenger based on these portions, allowing for targeted intrusion detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a monitoring system detects all human bodies in videos, then detection coverage is improved, but false alarms increase due to inability to distinguish pedestrians from vehicle occupants
Solution Approach 1:
The system segments the detection process into multiple stages: initial human body detection, portion detection (face, head, upper body, whole body), and state determination. This segmentation allows the system to analyze different aspects of detected objects separately, enabling accurate differentiation between pedestrians and vehicle occupants without generating false alarms.
Solution Approach 2:
The system applies different detection criteria and analysis methods to different portions of detected human bodies. By examining specific local features (such as the presence/absence of face, head, upper body portions) and their spatial relationships, the system can determine the state of each detected object, thereby improving detection accuracy while reducing false alarms from vehicle occupants.
2Reliability
If the system tracks all detected human bodies, then tracking coverage is improved, but processing efficiency decreases due to unnecessary tracking of vehicle occupants
Solution Approach 1:
The system performs preliminary state determination before initiating tracking. By analyzing detected portions and determining whether each detected human body is a pedestrian or vehicle occupant before tracking begins, the system avoids unnecessary tracking processing for vehicle occupants, thereby improving processing efficiency while maintaining comprehensive detection coverage.
Solution Approach 2:
The system extracts and processes only the necessary information for state determination (detection portions and their relationships) before tracking. By separating the state determination step from the tracking step and only applying tracking to pedestrians, the system reduces overall processing load while maintaining complete detection coverage.
3Device complexity
If the system uses simple human body detection, then device complexity is reduced, but detection precision decreases due to inability to differentiate object states
Solution Approach 1:
The detection system is segmented into modular components: a human body detection unit that identifies presence, a portion detection unit that identifies specific body parts (face, head, upper body, whole body), and a state determination unit that combines this information. This modular segmentation maintains relative system simplicity while enabling precise object state discrimination through systematic analysis of detection results.
Solution Approach 2:
The system uses a universal human body detection approach that can identify different body portions using the same detection framework. By detecting multiple portions (face, head, upper body, whole body) with consistent methods and then analyzing their combinations, the system achieves precise state discrimination without requiring entirely separate detection systems for each object type.
Data Source
AI summary
In order to detect an object by discriminating the object in accordance with the state of the object, a detection device comprises: a detection unit configured to detect a plurality of portions of an object contained in a frame image of a moving image; an intrusion determination unit configured to determine that the object has intruded into a preset area of the frame image; and a determination unit configured to determine whether to notify a determination result of the intrusion determination unit, in accordance with whether portions detected by the detection unit include a predetermined portion.