Faceprint Selection for Faster Mobile Face Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing facial recognition systems on mobile devices face challenges in accurately recognizing faces due to factors like brightness, positioning, and expression variations, especially when dealing with large image corpuses, leading to high computational complexity and time consumption.
Innovation Solution
A faceprint server optimizes facial recognition by generating and delivering optimized faceprints to client devices, reducing the number of comparisons needed by selecting a subset of users likely to appear in captured images, and using unique reference images compressed with less intensive calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If facial recognition compares all reference images in a large corpus, then recognition accuracy is improved, but computational complexity and time consumption increase significantly
Solution Approach 1:
The patent segments the large corpus of reference images into multiple subsets or groups. Instead of comparing the query image against all reference images simultaneously, the system divides the comparison task into smaller segments, reducing the computational burden on mobile devices while maintaining comprehensive coverage for accurate recognition.
Solution Approach 2:
The system performs preliminary actions by pre-processing and organizing reference images into structured subsets before the actual recognition task. This includes pre-computing features, organizing images by categories or attributes, and preparing optimized data structures that enable faster comparison during runtime without sacrificing accuracy.
2Measurement precision
If facial recognition compares all reference images in a large corpus, then recognition accuracy is improved, but time consumption increases significantly
Solution Approach 1:
The patent segments the large corpus of reference images into multiple subsets or groups. Instead of comparing the query image against all reference images simultaneously, the system divides the comparison task into smaller segments, reducing the computational burden on mobile devices while maintaining comprehensive coverage for accurate recognition.
Solution Approach 2:
The system performs preliminary actions by pre-processing and organizing reference images into structured subsets before the actual recognition task. This includes pre-computing features, organizing images by categories or attributes, and preparing optimized data structures that enable faster comparison during runtime without sacrificing accuracy.
3Device complexity
If optimized faceprints use a subset of users, then computational intensity is reduced, but recognition coverage may be limited
Solution Approach 1:
The system performs preliminary actions by pre-processing and organizing reference images into structured subsets before the actual recognition task. This includes pre-computing features, organizing images by categories or attributes, and preparing optimized data structures that enable faster comparison during runtime without sacrificing accuracy.
Solution Approach 2:
The patent implements dynamic adaptation where the system can adjust the scope and composition of reference image subsets based on the specific recognition task, device capabilities, and performance requirements. This allows the system to optimize computational intensity while maintaining adequate recognition coverage for different scenarios.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A server determines a plurality of faceprints representing a plurality of users to be recognized at a client device. Each faceprint contains a number of reference images for a given user that are used to recognize facial images of the user detecting in media captured at the client device. The faceprints delivered to the client device are determined for the client device based on the users likely to be detected in images captured at the client device. The reference images with a given faceprint delivered to the client device are selected by the server based on their recognition value in identifying the users likely to be detected in images captured at the client device.