Facial Recognition Using Social Network Affinity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional face recognition methods in social networking systems do not effectively utilize the wealth of information available through social networks, limiting their ability to provide accurate and relevant identification suggestions for users.
Innovation Solution
A social networking system that performs real-time facial recognition by detecting and tracking faces in images and videos, selecting candidates from the user's social network based on confidence levels and relationship coefficients, and incorporating these factors into an overall candidate score using a hidden Markov model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional face recognition methods are used, then the system can perform basic face detection, but the recognition accuracy and relevance are limited due to not utilizing social network information
Solution Approach 1:
The patent combines conventional face recognition technology with social network data by merging the detected face with candidate profiles from the social network. The system integrates visual recognition results with social relationship information, creating a unified identification system that leverages both technical detection and contextual social data to improve accuracy.
Solution Approach 2:
The patent introduces social network information as an intermediary layer between face detection and identification. Instead of directly identifying faces through recognition alone, the system uses social network data (connections, interactions, profiles) as a mediator to suggest and verify candidate identities, thereby enhancing recognition accuracy through additional contextual information.
2Measurement precision
If social network information is integrated into face recognition, then identification relevance improves, but system complexity increases
Solution Approach 1:
The patent segments the face recognition system into distinct functional modules: face detection module, candidate selection module (using social network data), scoring module, and identification module. This segmentation allows each component to handle specific tasks independently, managing system complexity through modular architecture while maintaining improved identification relevance.
Solution Approach 2:
The system performs preliminary actions by pre-selecting candidate profiles from the social network based on detected faces before final identification. This preliminary filtering using social connection data reduces the search space and simplifies the subsequent identification process, managing complexity by preparing information in advance rather than processing everything simultaneously.
3Productivity
If real-time facial recognition is performed on images and videos, then user experience improves, but processing time and computational resources increase
Solution Approach 1:
The patent applies partial action by not processing every detected face with full social network analysis. Instead, the system selectively applies detailed social network querying and candidate scoring only to faces that meet certain criteria (e.g., detection confidence threshold, user preferences), performing excessive action only where necessary to maintain real-time performance while reducing overall processing time.
Solution Approach 2:
The system uses periodic action by updating and refreshing social network data at intervals rather than continuously querying the social network for every face detection. This periodic refresh approach maintains real-time recognition capability while reducing computational overhead and processing time associated with constant social network access.
Data Source
AI summary
In particular embodiments, one or more images associated with a primary user are received. The image(s) may comprise single images, a series of related images, or video frames. In each image, one or more faces are detected and/or tracked. For each face, a set of one or more candidates are selected who may be identified with the face. A candidate score is calculated for each candidate based on a computed measure of affinity of the primary user for a particular candidate, a facial recognition score comparing the candidate to the face, and a geographic proximity of the candidate to the primary user at a time when the one or more images were created. A winning candidate is selected based on the candidate scores.