Research map / Architectures and efficiency / CNN design
MOL: Joint Estimation of Micro-Expression, Optical Flow, and Landmark via Transformer-Graph-Style Convolution
Zhiwen Shao, Yifan Cheng, Feiran Li, Yong Zhou, Xuequan Lu, Yuan Xie, Lizhuang Ma
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 optical flow estimation and landmark detection auxiliary tasks (ablation) | MOL · CASME II (5 classes) | accuracy: 66.55 -> 79.23 | Improved |
| add landmark detection task (ablation) | MOL · CASME II (5 classes) | accuracy: 76.60 -> 79.23 | Improved |
| add optical flow estimation task (ablation) | MOL · CASME II (5 classes) | accuracy: 75.48 -> 79.23 | Improved |
| add the F5C block (ablation) | MOL · CASME II (5 classes) | accuracy: 62.90 -> 79.23 | Improved |
| add fully-connected convolution (FCC) to F5C (ablation) | MOL · CASME II (5 classes) | accuracy: 65.45 -> 79.23 | Improved |
| add channel correspondence convolution (CCC) to F5C (ablation) | MOL · CASME II (5 classes) | accuracy: 76.86 -> 79.23 | Improved |
| stack 2 F5C blocks instead of 1 (ablation) | MOL · CASME II (5 classes) | accuracy: 79.23 -> 78.68 | Hurt |
| stack 3 F5C blocks instead of 1 (ablation) | MOL · CASME II (5 classes) | accuracy: 79.23 -> 78.27 | Hurt |
| use bidirectional FCC instead of vanilla transformer (ablation) | MOL · CASME II (5 classes) | accuracy: 76.30 -> 79.23 | Improved |
| use bidirectional FCC instead of one-directional FCC-V (ablation) | MOL · CASME II (5 classes) | accuracy: 78.04 -> 79.23 | Improved |
| concatenate consecutive frame features instead of element-wise addition (ablation) | MOL · CASME II (5 classes) | accuracy: 72.51 -> 79.23 | Improved |
| concatenate consecutive frame features instead of element-wise subtraction (ablation) | MOL · CASME II (5 classes) | accuracy: 72.13 -> 79.23 | Improved |
Closest papers by meaning
- From Pixels to Semantics: Unified Facial Action Representation Learning for Micro-Expression Analysis (ICLR 2026 · similarity 0.92)
- SelfME: Self-Supervised Motion Learning for Micro-Expression Recognition (CVPR 2023 · similarity 0.89)
- Micron-BERT: BERT-Based Facial Micro-Expression Recognition (CVPR 2023 · similarity 0.89)
- Region-Aware Instance Consistency Learning for Micro-Expression Recognition (CVPR 2026 · similarity 0.88)
- Revisiting the Volumetric Data of 4DME: Compression, Extension and Benchmarking for Micro-Expression Analysis (ECCV 2026 · similarity 0.88)
- CMNet: Contrastive Magnification Network for Micro-Expression Recognition (AAAI 2023 · similarity 0.88)
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.
