Face Selection Using Spatial Distance for Payment Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current face selection methods in offline devices for face-scanning payment scenarios face accuracy issues due to reliance on facial area size, which can be inconsistent and error-prone, especially when faces vary in size.
Innovation Solution
A method and device that select a target face based on spatial distance from the camera, using a combination of depth, horizontal, and vertical dimension distances, allowing for flexible selection criteria to improve accuracy by reducing the field of view and using deep learning for face detection and key point positioning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If face selection is based on largest facial area, then the selection process is simple, but the accuracy of selecting the target face deteriorates due to errors and face size variations
Solution Approach 1:
The patent transitions from 2D facial area measurement to 3D spatial distance measurement by introducing depth information. The system calculates spatial distance between the camera and each face using depth maps and key point positioning, thereby resolving the inaccuracy caused by relying solely on 2D area metrics that fail to account for distance variations.
Solution Approach 2:
The patent introduces depth maps and key point positioning technology as intermediary tools to bridge the gap between simple area-based selection and accurate target face identification. These intermediaries enable the system to calculate precise spatial distances without requiring complex multi-dimensional analysis, thus improving accuracy while maintaining operational simplicity.
2Ease of manufacture
If face selection is based on facial area size, then the method is easy to implement, but the reliability of face selection deteriorates due to large errors
Solution Approach 1:
The patent performs preliminary depth map acquisition and key point positioning before face selection. By pre-processing the spatial information and establishing reference points on faces, the system creates a reliable foundation for accurate target face selection, eliminating the need for complex real-time calculations during the selection process.
Solution Approach 2:
The patent changes the selection parameter from 2D facial area to 3D spatial distance. This parameter transformation fundamentally improves reliability by incorporating depth information, while the calculation method remains computationally efficient through the use of pre-acquired depth maps and key point coordinates.
Data Source
AI summary
Methods, systems, and devices, including computer programs encoded on computer storage media, for selecting a target face are provided. One of the methods includes: obtaining at least one facial area including one or more faces in an image taken by a camera; determining, based on the image, a spatial distance between each of the one or more faces and the camera; and selecting, based on the spatial distance, the target face from the one or more faces.