Research map / Architectures and efficiency / CNN design
BioNet: A Biologically-Inspired Network for Face Recognition
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 |
|---|---|---|---|
| Use BioNet with four observed attributes instead of GroupFace | ResNet-101 face recognition · IJB-B | TMR @1e-6: 52.12 -> 68.28 | Helped |
| Use BioNet with four observed attributes instead of GroupFace | ResNet-101 face recognition · IJB-C | TMR @1e-6: 89.28 -> 88.02 | Hurt |
| Add the CFC, CRT, and RIM modules together (ablation) | CASIA-Net face recognition · IJB-A | TMR @1e-4: 87.39 -> 89.31 | Helped |
| Add RIM to CFC and CRT (ablation) | CASIA-Net face recognition · IJB-A | TMR @1e-4: 87.97 -> 89.31 | Helped |
| Add CRT to CFC and RIM (ablation) | CASIA-Net face recognition · MegaFace | rank1@1e6: 87.65 -> 90.37 | Helped |
| Add CFC to CRT and RIM (ablation) | CASIA-Net face recognition · IJB-A | TMR @1e-4: 89.08 -> 89.31 | Helped |
| Add RIM to CFC and CRT (ablation) | ResNet-101 face recognition · IJB-C | TMR @1e-5: 86.36 -> 92.18 | Helped |
| Add CRT to CFC and RIM (ablation) | ResNet-101 face recognition · IJB-C | TMR @1e-5: 90.14 -> 92.18 | Helped |
| Add attribute supervision to BioNet (ablation) | ResNet-101 face recognition · IJB-C | TMR @1e-5: 88.05 -> 92.18 | Helped |
| Replace four latent attributes with four observed attributes (ablation) | ResNet-101 face recognition · IJB-C | TMR @1e-5: 90.66 -> 92.18 | Helped |
| Increase the number of observed attributes from 4 to 40 (ablation) | CASIA-Net face recognition · IJB-C | TMR @1e-5: 58.33 -> 69.46 | Helped |
| Increase the number of latent attributes from 4 to 40 (ablation) | CASIA-Net face recognition · IJB-C | TMR @1e-5: 55.32 -> 65.79 | Helped |
Closest papers by meaning
- Explicitly Modeling Subcortical Vision with a Neuro-Inspired Front-End Improves CNN Robustness (NeurIPS 2025 · similarity 0.86)
- SapiensID: Foundation for Human Recognition (CVPR 2025 · similarity 0.86)
- A Convolutional Neural Network Interpretable Framework for Human Ventral Visual Pathway Representation (AAAI 2024 · similarity 0.85)
- L-WISE: Boosting Human Visual Category Learning Through Model-Based Image Selection and Enhancement (ICLR 2025 · similarity 0.85)
- Convolution Goes Higher-Order: A Biologically Inspired Mechanism Empowers Image Classification (NeurIPS 2025 · similarity 0.84)
- Masked Face Recognition with Generative-to-Discriminative Representations (ICML 2024 · similarity 0.84)
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.