Research map / Vision-language and multimodal / Human and embodied modeling
Emotion Recognition in Context
Ronak Kosti, Jose M. Alvarez, Adria Recasens, Agata Lapedriza
Check 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 |
|---|---|---|---|
| Use body and whole-image inputs together instead of body input alone (ablation) | EMOTIC CNN · EMOTIC test | Average precision (AP): 25.44 -> 28.33 | Improved |
| Use body and whole-image inputs together instead of image input alone (ablation) | EMOTIC CNN · EMOTIC test | Average precision (AP): 22.48 -> 28.33 | Improved |
| Add whole-image context to body input (ablation) | EMOTIC CNN · EMOTIC test | Mean error rate: 1.0 -> 1.0 | No clear effect |
| Add whole-image context to image-only input (ablation) | EMOTIC CNN · EMOTIC test | Mean error rate: 1.2 -> 1.0 | Improved |
| Add continuous-dimension training loss alongside discrete-category loss (ablation) | Body + Image EMOTIC CNN · EMOTIC test | Average precision (AP): 24.18 -> 28.33 | Improved |
| Add discrete-category training loss alongside continuous-dimension loss (ablation) | Body + Image EMOTIC CNN · EMOTIC test | Mean error rate: 1.1 -> 1.0 | Improved |
| Use body and image together instead of body alone (ablation) | EMOTIC CNN · EMOTIC test, Valence | Mean error rate: 0.9 -> 0.9 | No clear effect |
| Use body and image together instead of image alone (ablation) | EMOTIC CNN · EMOTIC test, Valence | Mean error rate: 0.9 -> 0.9 | No clear effect |
| Use body and image together instead of body alone (ablation) | EMOTIC CNN · EMOTIC test, Arousal | Mean error rate: 1.1 -> 1.2 | Hurt |
| Use body and image together instead of image alone (ablation) | EMOTIC CNN · EMOTIC test, Arousal | Mean error rate: 1.9 -> 1.2 | Improved |
| Use body and image together instead of body alone (ablation) | EMOTIC CNN · EMOTIC test, Dominance | Mean error rate: 1.0 -> 0.9 | Improved |
| Use body and image together instead of image alone (ablation) | EMOTIC CNN · EMOTIC test, Dominance | Mean error rate: 0.8 -> 0.9 | Hurt |
Closest papers by meaning
- Context Based Emotion Recognition using EMOTIC Dataset (TPAMI 2019 · similarity 0.94)
- Emotion Recognition for Multiple Context Awareness (ECCV 2022 · similarity 0.90)
- EmoNet-Face: An Expert-Annotated Benchmark for Synthetic Emotion Recognition (NeurIPS 2025 · similarity 0.87)
- FindingEmo: An Image Dataset for Emotion Recognition in the Wild (NeurIPS 2024 · similarity 0.86)
- $E^3$: Exploring Embodied Emotion Through A Large-Scale Egocentric Video Dataset (NeurIPS 2024 · similarity 0.86)
- OV-MER: Towards Open-Vocabulary Multimodal Emotion Recognition (ICML 2025 · similarity 0.85)
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.
