Video Face Clustering for Accurate Age Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video sharing platforms face challenges in automatically detecting and preventing the upload of content featuring underage individuals, necessitating efficient age estimation methods to curb underage content and trafficking.
Innovation Solution
A system and method for age estimation involving facial detection, embedding, clustering, and classification of faces in videos or images, utilizing algorithms to identify and classify ages, with third-party confirmation for accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If automated age estimation systems are implemented to detect underage content, then the prevention of underage content upload is improved, but the complexity of the system increases
Solution Approach 1:
The system divides the age estimation process into distinct modules: facial detection, embedding generation, clustering, and age classification. Each module handles a specific task independently, making the overall complex system manageable and maintainable while improving reliability through specialized processing at each stage
Solution Approach 2:
The patent introduces an intermediary clustering step that groups facial embeddings before age classification. This intermediary layer simplifies the relationship between raw facial data and age estimates, reducing system complexity by creating a buffer layer that organizes data before final classification
2Measurement precision
If third-party age estimation is used for confirmation, then the measurement precision of age estimation is improved, but the loss of time increases
Solution Approach 1:
The system performs preliminary age classification using its own algorithms before seeking third-party confirmation. This preliminary action filters out cases that can be confidently classified internally, reserving third-party review only for ambiguous or critical cases, thereby reducing overall processing time while maintaining precision
Solution Approach 2:
The system applies third-party estimation selectively rather than universally. By using third-party confirmation only for specific cases where internal estimation is uncertain or when high precision is required, the system achieves improved measurement precision without the full time cost of universal third-party review
3Productivity
If facial detection and clustering algorithms are applied to video frames, then the productivity of age estimation is improved, but the use of energy by the system increases
Solution Approach 1:
The system extracts frames from video at periodic intervals rather than processing every frame continuously. This periodic sampling maintains productivity by processing sufficient frames for accurate age estimation while significantly reducing energy consumption by skipping frames between extractions
Solution Approach 2:
The patent combines multiple processing operations into unified algorithms that handle facial detection, embedding generation, and clustering in integrated operations. This merging reduces redundant computations and improves productivity while lowering energy consumption through more efficient algorithmic execution
Data Source
AI summary
Systems, methods, and computer-readable storage media for age estimation/classification, and more specifically to estimating/classifying the ages of people appearing within videos. Systems configured as disclosed herein can receive a video, then identify multiple facial images for each individual captured within the video. The system can create embeddings of the facial images, then cluster those images together based on distances between the corresponding embeddings. The system can also execute a matchability algorithm on those facial images, determining which of the images provides the clearest image of the individual(s), and can then estimate the age of the individual(s) using the best matchability images and/or send the best matchability image for each individual to a third party for analysis.