Face Recognition System Using Light Difference Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing facial recognition technologies in entrance guard and security systems are sensitive to light conditions and require fixed facial poses, leading to high computational complexity and delayed responses, making them unsuitable for real-time continuous detection.
Innovation Solution
A facial recognition system comprising a light pretreatment module using the difference of Gaussians method, a feature generation module employing local binary patterns, and a facial feature library module, capable of recognizing faces in various positions and light conditions, reducing computational complexity for real-time recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high computational complexity algorithms are used for facial recognition, then recognition accuracy is improved, but response time increases and real-time processing capability deteriorates
Solution Approach 1:
The patent segments the facial recognition process into multiple stages: light pretreatment module for initial image processing, feature generation module for extracting key characteristics, and feature matching module for comparison. This segmentation allows each module to perform specialized operations with optimized computational complexity, achieving real-time processing while maintaining accuracy.
Solution Approach 2:
The patent implements preliminary action by pre-processing facial images through light pretreatment before feature extraction. The light pretreatment module optimizes light difference degrees in advance, preparing the data in a form that reduces subsequent computational requirements during recognition, thereby enabling faster real-time response.
2Measurement precision
If fixed light source compensation is applied to improve recognition accuracy, then the system becomes sensitive to light conditions and requires fixed lighting, but adaptability to various light environments deteriorates
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting light difference degree optimization parameters in the light pretreatment module based on varying light conditions. Instead of requiring fixed lighting, the system adapts its processing parameters to different illumination environments, maintaining recognition accuracy across diverse light conditions including polarized and non-compensated light sources.
3Measurement precision
If fixed facial pose requirements are imposed to improve recognition accuracy, then the system can only recognize frontal views, but versatility in recognizing faces at different positions deteriorates
Solution Approach 1:
The patent implements universality by designing a facial recognition system that can handle multiple facial poses and positions through the same processing pipeline. The feature generation and matching modules are capable of extracting and comparing facial features regardless of orientation, enabling the system to recognize faces in various positions (frontal, profile, angled views) with consistent accuracy.
4Productivity
If continuous real-time facial detection is implemented, then response speed requirement increases, but computational resource consumption increases and existing systems cannot meet the requirement
Solution Approach 1:
The patent extracts only the essential and most discriminative facial features during the feature generation process, rather than processing all image data. By taking out only the critical feature elements needed for recognition, the system significantly reduces computational resource consumption while maintaining real-time processing capability and continuous detection performance.
Data Source
AI summary
The present invention relates to a face recognition system and a face recognition method, mainly comprising a lighting preprocessing module, a feature generation module, a facial feature library module, and a feature matching and recognition module. In the face recognition method, the inputted face image is first light-difference optimized by means of the lighting preprocessing module, then a feature vector of the face image is generated by means of the feature generation module; then, the feature vector to be recognized is matched with all of the feature vectors in the feature library and computed to arrive at an identity result corresponding to the feature vector to be recognized. In the technical solutions of the present invention, the identity of a face image to be recognized may be determined in real time for a system whose computing resources are limited, using a short period of time and having high accuracy.