Research map / Video / Video-language understanding
CLIP-AUTT: Test-Time Personalization with Action Unit Prompting for Fine-Grained Video Emotion Recognition
Muhammad Osama Zeeshan, Masoumeh Sharafi, Benoît Savary, Alessandro Lameiras Koerich, Marco Pedersoli, Eric Granger
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 minimum-entropy temporal window selection (ablation) | CLIP-AU · BioVid | WAR (%): 78.0 -> 78.7 | Improved |
| add minimum-entropy temporal window selection (ablation) | CLIP-AU · BioVid | F1 (%): 74.8 -> 75.2 | Improved |
| add AU prompt tuning at test time (ablation) | CLIP-AU · BioVid | WAR (%): 78.0 -> 80.0 | Improved |
| add AU prompt tuning at test time (ablation) | CLIP-AU · BioVid | F1 (%): 74.8 -> 76.4 | Improved |
| add both window selection and AU prompt tuning (ablation) | CLIP-AU · BioVid | WAR (%): 78.0 -> 81.5 | Improved |
| add both window selection and AU prompt tuning (ablation) | CLIP-AU · BioVid | F1 (%): 74.8 -> 78.0 | Improved |
| add test-time personalization (CLIP-AUTT) on top of CLIP-AU | CLIP-AU · StressID | F1 (%): 58.5 -> 77.9 | Improved |
| add test-time personalization (CLIP-AUTT) on top of CLIP-AU | CLIP-AU · BAH | WAR (%): 68.3 -> 69.8 | Improved |
| use AU prompts instead of AU-free fine-tuning (CLIP-AU vs full CLIP FT) | CLIP-ViT-B/32 · BioVid | WAR (%): 69.7 -> 78.0 | Improved |
| use AU prompts instead of AU-free fine-tuning (CLIP-AU vs EmoCLIP) | CLIP-AU · StressID | WAR (%): 67.0 vs 66.5 | Mixed |
| replace AU prompts with 46 generic class prompts (ablation) | CLIP-AUTT · StressID | WAR (%): nearly a 15% drop | Hurt |
| vary temporal window length from 4 to 72 frames (ablation) | CLIP-AUTT · BioVid | WAR (%) | No clear effect |
Closest papers by meaning
- Kronecker Mask and Interpretive Prompts are Language-Action Video Learners (ICLR 2025 · similarity 0.85)
- VidEmo: Affective-Tree Reasoning for Emotion-Centric Video Foundation Models (NeurIPS 2025 · similarity 0.84)
- AffectGPT: A New Dataset, Model, and Benchmark for Emotion Understanding with Multimodal Large Language Models (ICML 2025 · similarity 0.84)
- EmoVIT: Revolutionizing Emotion Insights with Visual Instruction Tuning (CVPR 2024 · similarity 0.84)
- ViLT-CLIP: Video and Language Tuning CLIP with Multimodal Prompt Learning and Scenario-Guided Optimization (AAAI 2024 · similarity 0.84)
- AUEditNet: Dual-Branch Facial Action Unit Intensity Manipulation with Implicit Disentanglement (CVPR 2024 · 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.
