Image Processing Apparatus Infrared Detection Printed Faces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional person detection methods using luminance and color information struggle to distinguish between actual people and faces in posters or monitors, as they incorrectly identify printed faces as actual individuals.
Innovation Solution
An image processing apparatus that captures both visible and invisible light images, using a dichroic mirror to separate wavelengths, and compares detection results between the two to accurately identify actual people by differentiating skin absorption patterns in infrared light from printed materials.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If luminance or color information is used to detect a face part, then person detection can be performed, but it becomes difficult to distinguish a face in a poster or monitor from an actual person
Solution Approach 1:
The patent transitions from two-dimensional visible light image analysis to three-dimensional spatial analysis by incorporating infrared light detection. The infrared camera detects heat radiation from actual persons, adding a thermal dimension that printed faces lack, thereby enabling accurate distinction between real and printed faces in monitoring systems
Solution Approach 2:
The patent introduces an intermediary infrared detection mechanism that mediates between visible light imaging and actual person identification. The infrared camera acts as an intermediary device that detects thermal signatures, providing additional verification layer to distinguish actual persons from printed faces without directly modifying the visible light detection process
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Effectively differentiates between actual people and printed faces, ensuring accurate detection of only the former in monitoring systems by leveraging the distinct absorption patterns in infrared light.
Implementation Method 1
captures both visible and invisible light images, using a dichroic mirror to separate wavelengths
Implementation Method 2
compares detection results between the two to accurately identify actual people by differentiating skin absorption patterns in infrared light from printed materials
Data Source
AI summary
An image processing apparatus detects a person in a visible light image obtained by capturing that uses visible light, detects a person in an invisible light image obtained by capturing that uses invisible light in a capturing region that includes a capturing region of the visible light image, and determines that a person detected in the visible light image who is not detected in the invisible light image is a person who is actually present.