Research map / Label-efficient and robust learning / Recognition and long-tail
The MegaFace Benchmark: 1 Million Faces for Recognition at Scale
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Increase the gallery from 10 to 1M distractors | Face recognition algorithms · FaceScrub and FG-NET | identification rate: above 95% -> 35-75% | Hurt |
| Increase the gallery from 10K to 1M distractors | Face recognition algorithms · FaceScrub | verification performance at low false accept rates: about 40% | Hurt |
| Use FG-NET rather than FaceScrub as the probe set | Face recognition algorithms · 1M-distractor verification | verification performance: about 60% | Hurt |
| Train on larger datasets | Face recognition algorithms · MegaFace identification | identification performance | Helped |
| Train FaceN on 18M photos rather than FaceNet on 500M photos | FaceN and FaceNet · FaceScrub verification | verification performance | No clear effect |
| Match photos with a larger age gap | Face recognition algorithms · FG-NET | rank-1 identification | Hurt |
| Recognize children rather than adults | Face recognition algorithms · FG-NET at scale | rank-1 identification | Hurt |
| Match photos with a larger pose difference | Face recognition algorithms · MegaFace identification | identification accuracy | Hurt |
| Scale the gallery from 10K to 1M distractors | Face recognition verification · FaceScrub and FG-NET | verification rate | No clear effect |
Closest papers by meaning
- Level Playing Field for Million Scale Face Recognition (CVPR 2017 · similarity 0.92)
- WebFace260M: A Benchmark Unveiling the Power of Million-Scale Deep Face Recognition (CVPR 2021 · similarity 0.89)
- HyperFace: Generating Synthetic Face Recognition Datasets by Exploring Face Embedding Hypersphere (ICLR 2025 · similarity 0.84)
- Vec2Face: Scaling Face Dataset Generation with Loosely Constrained Vectors (ICLR 2025 · similarity 0.84)
- AI-Face: A Million-Scale Demographically Annotated AI-Generated Face Dataset and Fairness Benchmark (CVPR 2025 · similarity 0.83)
- GIF: Generative Inspiration for Face Recognition at Scale (CVPR 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.