Concealed Item Detection via Video Motion Magnification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current security systems face inefficiencies in detecting concealed dangerous items, such as firearms or explosives, due to the need for radiographic screenings that are time-consuming and require individual screening, making them unsuitable for high-traffic areas.
Innovation Solution
A security system and method that uses a camera to capture image series of travelers in Eulerian or Lagrangian frames, processes these images to magnify small motions generated by concealed items, and activates an alarm system upon detection of characteristic kinematic behaviors indicative of concealed threats.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If radiographic screenings are used to detect concealed items, then detection reliability is improved, but screening time and operational complexity increase
Solution Approach 1:
The patent replaces radiographic screening systems with a computer vision-based detection system that uses standard cameras and image processing algorithms. This substitution eliminates the need for heavy radiographic machinery while maintaining detection capability through analysis of clothing texture patterns, contours, and motion characteristics in video footage.
Solution Approach 2:
The system creates a visual copy or representation of the concealed item's presence through image processing techniques. By analyzing and enhancing visual data from cameras, the system generates detectable patterns that indicate concealed items without requiring direct physical or radiographic interaction with the subject.
2Reliability
If radiographic screenings are used to detect concealed items, then detection reliability is improved, but device complexity and cost increase
Solution Approach 1:
The patent replaces complex radiographic screening systems with a computer vision-based detection system that uses standard cameras and image processing algorithms. This substitution eliminates the need for heavy radiographic machinery while maintaining detection capability through analysis of clothing texture patterns, contours, and motion characteristics in video footage.
Solution Approach 2:
The system uses standard cameras and general-purpose computing devices that can perform multiple functions, rather than specialized radiographic equipment. The same camera system can be used for various surveillance and security applications, reducing overall system complexity and cost while maintaining detection effectiveness.
3Reliability
If radiographic screenings are used to detect concealed items, then detection capability is improved, but ease of operation deteriorates due to individual screening requirements
Solution Approach 1:
The system enables continuous monitoring and detection of concealed items as individuals move through the monitored area. Unlike radiographic screening that requires stopping each person, this system maintains constant surveillance and automatically processes video feeds in real-time, allowing uninterrupted flow while sustaining detection capability.
Solution Approach 2:
The system automatically performs detection without requiring operators to manually screen each individual. The computer vision algorithms autonomously analyze video footage, identify suspicious patterns, and generate alerts, reducing operational complexity and making the system easier to deploy and maintain.
Data Source
AI summary
A method to analyze video includes obtaining an image series in an Eulerian or Lagrangian frame of reference, selecting one or more specific regions of interest and extracting to remove irrelevant motion and/or noise, decomposing the extracted images into a plurality of frequency bands and extracting a pixel value time series corresponding to the values of a pixel in each spatial frequency band, magnifying the pixel value time series to obtain a magnified pixel value time series, adding the magnified pixel value time series to the pixel value time series to generate a superimposed pixel value time series, and applying a spatial reconstruction to the superimposed pixel value time series to generate an output image series. A system to perform this method is also provided.