VTF900: A 900-Subject Visible-Thermal Face Recognition Dataset and Benchmark

Key points

• Despite advancements in visible-thermal face recognition, where thermal probe (query) face images are matched against visible enrollment templates, existing benchmarks are currently limited by small gallery sizes (N $\approx$ 100 subjects).1
• VTF900 is designed to support large-scale (i.e., large gallery) evaluations across verification (1:1 matching) and closed-set identification (1:N matching) for visible-thermal face recognition.1
• In this paper, after training models on the ARL-VTF dataset, we provide benchmark analyses across convolutional (baseline), vision transformers, and domain-adaptive neural network architectures to assess how these methods scale.1
• VTF900 exposes these limitations, enabling more rigorous evaluation of cross-spectral face recognition systems.1

Sources (1)

• [1]VTF900: A 900-Subject Visible-Thermal Face Recognition Dataset and Benchmark
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 17, 07:27 PM

  Therefore, we introduce VTF900, a large-scale visible-thermal face recognition benchmark with 900 subjects and paired imagery that includes expression, pose, and occlusion variations.

  Despite advancements in visible-thermal face recognition, where thermal probe (query) face images are matched against visible enrollment templates, existing benchmarks are currently limited by small gallery sizes (N $\approx$ 100 subjects).

Extractive summary: sentences quoted from the sources.