Research map / Label-efficient and robust learning / Recognition and long-tail
Recognizability Embedding Enhancement for Very Low-Resolution Face Recognition and Quality Estimation
Read the paperCheck 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 |
|---|---|---|---|
| add index diversion loss (ablation) | ArcFace baseline · TinyFace without distractors | Rank-1 IR (%): 70.333 -> 71.298 | Helped |
| add perceptibility attention loss (ablation) | ArcFace baseline · TinyFace without distractors | Rank-1 IR (%): 70.333 -> 71.540 | Helped |
| add RI regression loss (ablation) | ArcFace baseline · TinyFace without distractors | Rank-1 IR (%): 70.333 -> 71.674 | Helped |
| combine index diversion, attention, and RI regression losses (ablation) | ArcFace baseline · TinyFace without distractors | Rank-1 IR (%): 70.333 -> 71.915 | Helped |
| increase the RI regression loss weight above 10 (ablation) | Full method · TinyFace | convergence | Hurt |
| increase the attention MSE loss weight (ablation) | Full method · TinyFace | overall performance | Hurt |
| use recognizability-aware embedding learning instead of the reproduced AdaFace baseline | ResNet-50 · TinyFace with distractors | Rank-1 IR (%): 71.38 -> 73.06 | Helped |
| use recognizability-aware embedding learning instead of the reproduced AdaFace baseline | ResNet-50 · SurvFace at FAR 0.001 | TPR (%): 40.09 -> 48.48 | Helped |
| use recognizability-aware embedding learning instead of the reproduced AdaFace baseline | ResNet-50 · SCFace 1.0m | Rank-1 IR (%): 99.84 -> 99.80 | Hurt |
Closest papers by meaning
- SER-FIQ: Unsupervised Estimation of Face Image Quality Based on Stochastic Embedding Robustness (CVPR 2020 · similarity 0.87)
- SDD-FIQA: Unsupervised Face Image Quality Assessment With Similarity Distribution Distance (CVPR 2021 · similarity 0.86)
- Representation Rigidity in Face Embeddings: Orthogonal Identifiability and Backfill-Free Compatibility (NeurIPS 2026 · similarity 0.85)
- Masked Face Recognition with Generative-to-Discriminative Representations (ICML 2024 · similarity 0.83)
- Vec2Face: Scaling Face Dataset Generation with Loosely Constrained Vectors (ICLR 2025 · similarity 0.83)
- Attributes Shape the Embedding Space of Face Recognition Models (ICML 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.