Learning from a single labeled face and a stream of unlabeled data

arXiv cs.LG / 5/1/2026

📰 NewsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper studies face recognition in a highly constrained setting where only one labeled image per (single) enrolled person is available and no labeled data exist for others.
It frames the task as one-class classification, addressing the additional difficulty that the model lacks negative examples.
The authors propose a non-parametric approach that learns from one labeled face image combined with a continuous stream of unlabeled data collected in realistic interaction scenarios.
Experiments on a dataset of 43 people show the method can recognize the target person about 90% of the time with nearly zero false positives, improving recall by 25% or more over the best baseline.
The work includes a sensitivity analysis and practical guidelines for tuning algorithm parameters.

Abstract

Face recognition from a single image per person is a challenging problem because the training sample is extremely small. We consider a variation of this problem. In our problem, we recognize only one person, and there are no labeled data for any other person. This setting naturally arises in authentication on personal computers and mobile devices, and poses additional challenges because it lacks negative examples. We formalize our problem as one-class classification, and propose and analyze an algorithm that learns a non-parametric model of the face from a single labeled image and a stream of unlabeled data. In many domains, for instance when a person interacts with a computer with a camera, unlabeled data are abundant and easy to utilize. This is the first paper that investigates how these data can help in learning better models in the single-image-per-person setting. Our method is evaluated on a dataset of 43 people and we show that these people can be recognized 90% of time at nearly zero false positives. This recall is 25+% higher than the recall of our best performing baseline. Finally, we conduct a comprehensive sensitivity analysis of our algorithm and provide a guideline for setting its parameters in practice.