Open-Set Face Identification on Few-Shot Gallery by Fine-Tuning
This work addresses a realistic face identification problem for security or surveillance applications where only limited enrollment data is available, representing an incremental improvement over existing methods.
The paper tackles open-set face identification with few-shot galleries by proposing a fine-tuning scheme with classifier weight imprinting and exclusive BatchNorm tuning, along with a Neighborhood Aware Cosine matcher, achieving improved rejection accuracy on unknown identities as validated on large-scale benchmarks.
In this paper, we focus on addressing the open-set face identification problem on a few-shot gallery by fine-tuning. The problem assumes a realistic scenario for face identification, where only a small number of face images is given for enrollment and any unknown identity must be rejected during identification. We observe that face recognition models pretrained on a large dataset and naively fine-tuned models perform poorly for this task. Motivated by this issue, we propose an effective fine-tuning scheme with classifier weight imprinting and exclusive BatchNorm layer tuning. For further improvement of rejection accuracy on unknown identities, we propose a novel matcher called Neighborhood Aware Cosine (NAC) that computes similarity based on neighborhood information. We validate the effectiveness of the proposed schemes thoroughly on large-scale face benchmarks across different convolutional neural network architectures. The source code for this project is available at: https://github.com/1ho0jin1/OSFI-by-FineTuning