Living Body Detection via Depth Difference Histograms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Three-dimensional face living body detection methods face challenges in efficiency, accuracy, and stability due to long curvature calculation times and limited use of global information, as well as reliance on threshold-based differentiation between living and non-living bodies.
Innovation Solution
A method involving the extraction of valid depth data from depth map data, generation of a depth difference histogram, and inputting it into a pre-trained machine learning classifier to determine if a target object is a living body, eliminating the need for curvature calculation and improving detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If curvature calculation is used for three-dimensional face living body detection, then detection accuracy can be improved, but calculation time increases significantly and global information is wasted
Solution Approach 1:
The patent extracts only the necessary depth information from depth map data rather than performing complete curvature calculations. By extracting valid depth data and generating depth difference histograms, the method obtains sufficient detection features without the computational overhead of full curvature analysis, thus improving efficiency while maintaining accuracy.
Solution Approach 2:
The patent segments the depth map data into valid depth data through image segmentation based on preset depth difference ranges. This segmentation approach processes only relevant portions of the data rather than performing global curvature calculations, reducing computational time while preserving essential depth characteristics for accurate detection.
2Device complexity
If threshold-based differentiation is used to distinguish living and non-living bodies, then the detection process is simple, but detection accuracy is limited
Solution Approach 1:
The patent transforms the detection approach from simple threshold comparison to machine learning classification based on depth difference histograms. By changing the parameter representation from raw depth values to histogram features and using pre-trained classifiers, the system achieves higher accuracy while maintaining relatively simple implementation through feature transformation.
3Productivity
If only depth information near feature points is extracted for curvature calculation, then calculation efficiency is improved, but global information of the three-dimensional face image is wasted
Solution Approach 1:
The patent creates depth difference histograms that serve multiple purposes: they capture local depth variations near feature points for efficient processing, while also preserving global depth distribution characteristics through histogram binning. This multi-functional representation maintains both calculation efficiency and global information utilization.
4Measurement precision
If machine learning classification is used for living body detection, then detection accuracy is improved, but algorithm overhead increases
Solution Approach 1:
The patent uses pre-trained machine learning classifiers that have already learned optimal detection patterns during offline training. During actual detection, the pre-trained models directly classify depth difference histograms without requiring complex real-time training or feature engineering, reducing online algorithm overhead while maintaining high accuracy through previously learned knowledge.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
A living body detection method and device are disclosed. Wherein the method comprises the following steps: extracting valid depth data of a target detection object from depth map data containing the target detection object; generating a depth difference histogram based on the valid depth data; and inputting the depth difference histogram into a pre-trained machine learning classifier to obtain a determination result of whether the target detection object is a living body. By adopting this method, the detection accuracy can be improved.