Distributed Identity Determination System for Privacy-Safe Face Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face recognition systems for passenger flow statistics in shopping malls face challenges with high computing resource demands and network load, leading to instability and accuracy issues, as well as concerns over user privacy when uploading video streams to back-end servers.
Innovation Solution
A distributed identity determination system where computing nodes collect images, perform face recognition, and upload facial features and node identity information to a server, enabling efficient and secure identity determination through distributed computation, with features like preset image collecting frequencies, worker threads for parallel processing, and local storage to manage data transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If video streams are uploaded to back-end servers for face recognition, then identity determination can be performed, but network load increases and system stability deteriorates
Solution Approach 1:
The system divides the face recognition processing into segments: computing nodes extract facial features and upload only feature data, while the back-end server performs identity matching. This segmentation reduces network load and improves system stability while maintaining accuracy.
Solution Approach 2:
The patent extracts only essential facial feature data from complete video streams for upload to the server. This extraction approach maintains identity determination accuracy while significantly reducing network bandwidth consumption and improving system reliability.
2Measurement precision
If complete video streams are uploaded for face recognition, then accurate identity determination is achieved, but user privacy is compromised
Solution Approach 1:
The system extracts only facial feature data from complete video streams, uploading minimal necessary information to the server. This preserves user privacy by preventing exposure of sensitive video content while maintaining sufficient data for accurate identity determination.
Solution Approach 2:
Facial feature extraction acts as an intermediary that transforms sensitive video data into anonymized feature representations. This mediator preserves identity information needed for recognition while removing personally identifiable visual content, thus protecting user privacy.
3Measurement precision
If face recognition processing is performed on all uploaded data, then accurate facial feature recognition is achieved, but computing resource consumption increases
Solution Approach 1:
The computing nodes perform preliminary extraction of facial features from video frames before upload. This pre-extraction reduces the amount of data requiring intensive processing at the server, lowering overall computing resource consumption while preserving recognition accuracy.
Solution Approach 2:
Facial feature extraction is performed as a preliminary action at the computing node before data transmission. This advance processing reduces the computational burden on the back-end server, optimizing resource allocation while maintaining recognition precision.
Data Source
AI summary
An identity determination system and an identity determination method are provided. The system includes: a computing node including an image acquisition device, and a server in communication connection with the computing node. The computing node is configured to: collect, by using the image acquisition device, an image according to a preset image collecting frequency; perform face recognition on the collected image to recognize a facial feature in the image; determine node identity information corresponding to the recognized facial feature according to the recognized facial feature; and upload the facial feature and the node identity information corresponding to the facial feature to the server. The server is configured to determine an identity of the facial feature based on the facial feature and the node identity information corresponding to the facial feature uploaded by the computing node.