Device and method for generating a biometric image of a person's face
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current biometric image capturing systems require multiple locations for photographing and application, are space-intensive, and face ambiguous user guidance due to language barriers, making the process inefficient and time-consuming.
Innovation Solution
A compact device using a semi-transparent mirror with integrated display and camera, enabling augmented reality-guided user instructions for capturing biometric images, suitable for confined spaces and accommodating diverse user needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 3D sensors (time of flight, structured light, ultrasonic) are used to capture accurate depth information, then depth map accuracy is improved, but device complexity and cost increase
Solution Approach 1:
The patent introduces an intermediary computational process (depth estimation algorithm) that mediates between the simple 2D image capture and the desired 3D depth information. Instead of using complex 3D sensors, the system uses a 2D image processor to generate depth maps through image processing techniques, thereby achieving depth information without the complexity of dedicated 3D sensing hardware
Solution Approach 2:
The patent creates a computational copy of depth information by generating synthetic depth maps from 2D images. Rather than directly capturing physical depth data with complex sensors, the system creates an equivalent representation through image processing algorithms, effectively copying the depth information that would otherwise require specialized sensors to obtain
2Manufacturing precision
If multiple image processing techniques (grayscale conversion, edge detection, depth estimation) are applied sequentially, then face image quality is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-processing images with multiple techniques (grayscale conversion, edge detection, depth estimation) before the actual face recognition task. By preparing enhanced depth maps and processed images in advance, the system reduces the computational burden during real-time recognition, thereby reducing overall processing time despite the additional preprocessing steps
Solution Approach 2:
The patent segments the image processing task into distinct independent stages (color to grayscale conversion, edge detection, depth map generation, face region extraction). Each segment can be processed independently and optimized separately, allowing for more efficient computation and parallel processing where applicable, thus managing processing time while maintaining image quality
3Loss of information
If face recognition is performed on original color images, then color information is preserved, but processing complexity and false recognition risk increase
Solution Approach 1:
The patent extracts only the essential features (edge information, depth information, grayscale intensity) from the original color images while discarding the color component. By taking out and removing the color information that causes processing complexity and false recognition, the system achieves more reliable and accurate face recognition while still capturing the critical structural features needed for identification
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a device (100) for generating a biometric image of a person's face using a semi-transparent mirror (103), the surface of which reflects on the front side (104) and transmits on the rear side (109), which is opposite the front side (104), having at least one display device (110) which is arranged on the rear side (109) of the semi-transparent mirror (103) and which is directed at the transmitting surface of the rear side (109) of the semi-transparent mirror (103), and having at least one camera (101) for capturing an image of the person's face, the beam path of which camera extends through the semi-transparent mirror (103) to the person's face, and having a control unit (102) and an output unit (111) which are designed to output information or an instruction as visual or audible user guidance for generating an augmented reality. The invention further relates to a method for generating a biometric image.