Research map / Label-efficient and robust learning / Recognition and long-tail
Exploring Disentangled Feature Representation Beyond Face Identification
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 |
|---|---|---|---|
| disentangle features into identity-distilled and complementary identity-dispelled streams | D²AE trained on WebFace · LFW | accuracy: 98.93 -> 99.25 | Helped |
| disentangle features into identity-distilled and complementary identity-dispelled streams | D²AE trained on WebFace · LFW | TPR@0.001FPR: 94.87 -> 96.80 | Helped |
| use the combined two-stream representation instead of the identity-distilling branch alone (ablation) | D²AE · LFW | identity accuracy: 99.78 -> 99.80 | No clear effect |
| use the combined two-stream representation instead of the identity-distilling branch alone (ablation) | D²AE · LFW | TPR@0.001FPR: 99.63 -> 99.40 | Hurt |
| add the identity-dispelling branch to the identity-distilling branch (ablation) | D²AE · CelebA and LFWA | attribute accuracy: 79.78 -> 83.16 | Helped |
| add identity confusion loss L_H (ablation) | D²AE-P · LFW | identity accuracy: 71.2 -> 64.13 | Hurt |
| add identity confusion loss L_H (ablation) | D²AE-P · CelebA and LFWA | attribute accuracy: 80.47 -> 81.99 | Helped |
| add adversarial identity classification loss L_I^adv (ablation) | D²AE-P · LFW | identity accuracy: 67.13 -> 64.13 | Hurt |
| add adversarial identity classification loss L_I^adv (ablation) | D²AE-P · CelebA and LFWA | attribute accuracy: 78.32 -> 81.99 | Helped |
| combine identity-distilled and identity-dispelled features (ablation) | D²AE · CelebA and LFWA | attribute accuracy: 81.99 -> 83.16 | Helped |
Closest papers by meaning
- 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 0.82)
- DIFFER: Disentangling Identity Features via Semantic Cues for Clothes-Changing Person Re-ID (CVPR 2025 · similarity 0.82)
- Privacy-Preserving Face Recognition Using Trainable Feature Subtraction (CVPR 2024 · similarity 0.82)
- Disguise without Disruption: Utility-Preserving Face De-identification (AAAI 2024 · similarity 0.82)
- FaceA-Net: Facial Attribute-Driven ID Preserving Image Generation Network (AAAI 2025 · similarity 0.82)
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.