Research map / Vision-language and multimodal / CLIP and zero-shot
General Facial Representation Learning in a Visual-Linguistic Manner
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 |
|---|---|---|---|
| Add 1-layer masked image modeling to image-text contrastive pretraining (ablation) | FaRL pretraining · LaPa | F1-mean: 91.75 -> 91.82 | Helped |
| Add 1-layer masked image modeling to image-text contrastive pretraining (ablation) | FaRL pretraining · AFLW-19 | NME_diag: 1.009 -> 1.004 | Helped |
| Add 1-layer masked image modeling to image-text contrastive pretraining (ablation) | FaRL pretraining · CelebA | mAcc: 91.31 -> 91.22 | Hurt |
| Add aligned face crops to ITC+MIM1 (ablation) | FaRL pretraining · LaPa | F1-mean: 91.82 -> 92.32 | Helped |
| Add aligned face crops to ITC+MIM1 (ablation) | FaRL pretraining · AFLW-19 | NME_diag: 1.004 -> 0.991 | Helped |
| Add aligned face crops to ITC+MIM1 (ablation) | FaRL pretraining · CelebA | mAcc: 91.22 -> 91.39 | Helped |
| Use a 6-layer rather than a 1-layer MIM head (ablation) | FaRL pretraining with alignment · LaPa | F1-mean: 92.32 -> 92.19 | Hurt |
| Use a 6-layer rather than a 1-layer MIM head (ablation) | FaRL pretraining with alignment · AFLW-19 | NME_diag: 0.991 -> 1.002 | Hurt |
| Use a 6-layer rather than a 1-layer MIM head (ablation) | FaRL pretraining with alignment · CelebA | mAcc: 91.39 -> 91.38 | No clear effect |
| Pretrain on LAION-FACE rather than equally sized randomly sampled LAION data (ablation) | ITC pretraining · CelebA | mAcc: 90.76 -> 91.31 | Helped |
| Pretrain on LAION-FACE rather than equally sized randomly sampled LAION data (ablation) | ITC pretraining · LaPa | F1-mean: 91.68 -> 91.75 | No clear effect |
| Pretrain on LAION-FACE rather than equally sized randomly sampled LAION data (ablation) | ITC pretraining · AFLW-19 | NME_diag: 1.010 -> 1.009 | No clear effect |
Closest papers by meaning
- Self-Supervised Facial Representation Learning with Facial Region Awareness (CVPR 2024 · similarity 0.87)
- FSFM: A Generalizable Face Security Foundation Model via Self-Supervised Facial Representation Learning (CVPR 2025 · similarity 0.85)
- HyperFace: Generating Synthetic Face Recognition Datasets by Exploring Face Embedding Hypersphere (ICLR 2025 · similarity 0.85)
- Masked Face Recognition with Generative-to-Discriminative Representations (ICML 2024 · similarity 0.85)
- SegFace: Face Segmentation of Long-Tail Classes (AAAI 2025 · similarity 0.84)
- FaceBench: A Multi-View Multi-Level Facial Attribute VQA Dataset for Benchmarking Face Perception MLLMs (CVPR 2025 · 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.