Anonymous Facial Recognition via Local Embedding Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional facial recognition systems compromise privacy by storing identity information and are vulnerable to malicious attacks, as they require pre-training and struggle with adaptiveness in different environments, leading to inefficient classification of known versus unknown individuals.
Innovation Solution
A novel framework that uses anonymous facial recognition, processing images locally to send only coded face embeddings for comparison, storing data in a 'gallery' without identifying information, allowing real-time adaptation and dynamic classification of individuals based on frequency and recency of visits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional facial recognition systems store identity information for recognition, then recognition accuracy is improved, but privacy security deteriorates
Solution Approach 1:
The system extracts only the essential facial features (face embeddings) from complete facial images, storing only these coded representations rather than full images or identity information. This extraction principle enables recognition functionality while removing unnecessary personal data that would compromise privacy.
Solution Approach 2:
The patent introduces face embeddings as an intermediary representation between the captured facial image and the recognition database. These embeddings serve as a mediator that preserves recognition accuracy while preventing direct access to identifiable facial data, thus protecting privacy security.
2Adaptability or versatility
If conventional systems pre-label and store identity data, then identification capability is improved, but system vulnerability to attacks worsens
Solution Approach 1:
The system uses temporary, non-persistent facial embeddings that are not permanently stored or backed up in vulnerable databases. These embedding representations are designed to be ephemeral, reducing the attack surface and potential impact of malicious attacks while maintaining identification capability during active operation.
3Measurement precision
If systems capture and transmit full images for processing, then analysis accuracy is improved, but data transmission security worsens
Solution Approach 1:
The system extracts only the essential facial features (face embeddings) from complete facial images, storing only these coded representations rather than full images or identity information. This extraction principle enables recognition functionality while removing unnecessary personal data that would compromise privacy.
Solution Approach 2:
The patent applies local quality by processing and extracting features directly at the client device rather than transmitting full images to remote servers. This localized processing ensures that only minimal coded data (face embeddings) are transmitted over the network, enhancing data transmission security while maintaining analysis accuracy through client-side computation.
Data Source
AI summary
Disclosed are systems and methods for improving interactions with and between computers in computerized security and content monitoring, hosting and providing devices, systems and/or platforms. The disclosed systems and methods provide a novel framework that adaptively distinguishes between known people versus unknown people based on a dynamically applied, anonymous facial recognition methodology. The disclosed framework provides such functionality by recognizing faces within captured images without storing any information or annotations regarding or revealing the captured person's identity. The framework is configured to adaptively learn to distinguish between faces seen for the first time and faces it has previously seen by locally processing a captured image and only sending face embeddings to a network location for future comparisons of subsequently, anonymously captured images.