Research map / Label-efficient and robust learning / Self-supervised and contrastive
SelfME: Self-Supervised Motion Learning for Micro-Expression 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 self-supervised learned motion with the full SelfME framework instead of FeatRef's TV-L1 input | SelfME · CASME II | UAR: 0.8873 -> 0.9290 | Helped |
| Use self-supervised learned motion with the full SelfME framework instead of FeatRef's TV-L1 input | SelfME · SMIC-HS | UF1: 0.7011 -> 0.6972 | Hurt |
| Add symmetric contrastive learning (ablation) | SelfME motion classifier · CASME II | UF1: 0.8468 -> 0.8903 | Helped |
| Add B-spline transformation (ablation) | SelfME motion classifier · CASME II | UF1: 0.8468 -> 0.8629 | Helped |
| Add motion amplification (ablation) | SelfME motion classifier · CASME II | UF1: 0.8468 -> 0.8718 | Helped |
| Use B-spline transformation and SC without MA (ablation) | SelfME · CASME II | UF1: 0.8903 -> 0.8923 | Helped |
| Add MA to B-spline transformation and SC (ablation) | SelfME · CASME II | UF1: 0.8923 -> 0.9078 | Helped |
| Set motion amplification to ×2 rather than ×1 (ablation) | SelfME · SMIC-HS | UF1: 0.6768 -> 0.6972 | Helped |
| Increase motion amplification from ×2 to ×3 (ablation) | SelfME · SMIC-HS | UF1: 0.6972 -> 0.6523 | Hurt |
| Replace SelfME motion with TV-L1 motion (ablation) | ViT classifier · CASME II | UF1: 0.8060 vs 0.8784 | Hurt |
| Replace SelfME motion with TCAE motion (ablation) | SCViT classifier · CASME II | UF1: 0.6158 vs 0.9078 | Hurt |
Closest papers by meaning
- From Pixels to Semantics: Unified Facial Action Representation Learning for Micro-Expression Analysis (ICLR 2026 · similarity 0.88)
- PrefAce: Face-Centric Pretraining with Self-Structure Aware Distillation (AAAI 2024 · similarity 0.86)
- FSRT: Facial Scene Representation Transformer for Face Reenactment from Factorized Appearance Head-pose and Facial Expression Features (CVPR 2024 · similarity 0.84)
- SMILE: Infusing Spatial and Motion Semantics in Masked Video Learning (CVPR 2025 · similarity 0.83)
- Self-Supervised Facial Representation Learning with Facial Region Awareness (CVPR 2024 · similarity 0.83)
- X-NeMo: Expressive Neural Motion Reenactment via Disentangled Latent Attention (ICLR 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.