Research map / Video / Video segmentation and tracking
CMNet: Contrastive Magnification Network 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 |
|---|---|---|---|
| Add intensity distillation loss (LID) to the classifier baseline (ablation) | CMNet · CASME II | Acc: 63.82 -> 73.58 | Helped |
| Add intensity distillation loss (LID) to the classifier baseline (ablation) | CMNet · SAMM | Acc: 62.50 -> 71.32 | Helped |
| Add intensity distillation loss (LID) to the classifier baseline (ablation) | CMNet · SMIC-HS | Acc: 65.85 -> 74.39 | Helped |
| Add Wilcoxon rank sum tendency calibration (LW rst) alongside LID (ablation) | CMNet · CASME II | Acc: 73.58 -> 78.05 | Helped |
| Add Wilcoxon rank sum tendency calibration (LW rst) alongside LID (ablation) | CMNet · SAMM | Acc: 71.32 -> 78.68 | Helped |
| Add Wilcoxon rank sum tendency calibration (LW rst) alongside LID (ablation) | CMNet · SMIC-HS | Acc: 74.39 -> 79.27 | Helped |
| Add LW rst without intensity distillation loss (ablation) | CMNet · CASME II | Acc: 63.82 -> 65.45 | Helped |
| Add LW rst without intensity distillation loss (ablation) | CMNet · SAMM | Acc: 62.50 -> 63.97 | Helped |
| Add LW rst without intensity distillation loss (ablation) | CMNet · SMIC-HS | Acc: 65.85 -> 65.85 | No clear effect |
| Use probabilistic rather than deterministic negative sampling (ablation) | CMNet · CASME II, SAMM, and SMIC-HS | Acc: 74.80 -> 78.05; 75.74 -> 78.68; 78.05 -> 79.27 | Helped |
| Use probabilistic rather than random negative sampling (ablation) | CMNet · CASME II, SAMM, and SMIC-HS | Acc: 75.61 -> 78.05; 76.47 -> 78.68; 76.22 -> 79.27 | Helped |
| Use probabilistic negative sampling instead of deterministic or random sampling (ablation) | CMNet · CASME II, SAMM, and SMIC-HS | F1-score: 70.30 -> 73.99; 73.94 -> 77.19; 78.35 -> 80.11 | Helped |
Closest papers by meaning
- Region-Aware Instance Consistency Learning for Micro-Expression Recognition (CVPR 2026 · similarity 0.87)
- A unique M-pattern for micro-expression spotting in long videos (ICLR 2024 · similarity 0.86)
- AUEditNet: Dual-Branch Facial Action Unit Intensity Manipulation with Implicit Disentanglement (CVPR 2024 · similarity 0.85)
- Prototypical Calibrating Ambiguous Samples for Micro-Action Recognition (AAAI 2025 · similarity 0.83)
- TEASER: Token Enhanced Spatial Modeling for Expressions Reconstruction (ICLR 2025 · similarity 0.82)
- Trend-Aware Supervision: On Learning Invariance for Semi-supervised Facial Action Unit Intensity Estimation (AAAI 2024 · 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.