Research map / Vision-language and multimodal / Multimodal LLMs
ContextFace: Generating Facial Expressions from Emotional Contexts
Min-jung Kim, Minsang Kim, Seung Jun Baek
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 Expression Generation dataset to training (ablation) | ContextFace · SFEW-C emotion recognition (quote) | UAR: 68.26 -> 79.34 | Improved |
| add Situation Generation dataset to training (ablation) | ContextFace · SFEW-C emotion recognition (quote) | UAR: 68.26 -> 88.33 | Improved |
| add both Situation and Expression Generation datasets (ablation) | ContextFace · SFEW-C emotion recognition (quote) | UAR: 68.26 -> 93.21 | Improved |
| add Expression Generation dataset to Situation dataset (ablation) | ContextFace · SFEW-C emotion recognition (quote) | UAR: 88.33 -> 93.21 | Improved |
| add Situation Generation dataset to Expression dataset (ablation) | ContextFace · SFEW-C emotion recognition (quote) | UAR: 79.34 -> 93.21 | Improved |
| scale backbone from LLaVA-llama2-7B to 13B (ablation) | ContextFace · SFEW-C emotion recognition (quote) | UAR: 89.71 -> 93.21 | Improved |
| scale backbone from LLaVA-llama2-7B to 13B (WAR) (ablation) | ContextFace · SFEW-C emotion recognition (quote) | WAR: 90.56 -> 94.89 | Improved |
| use ContextFace instead of Qwen2.5-VL-7B (situation) | ContextFace · SFEW-C zero-shot with situation | UAR: 86.08 -> 90.13 | Improved |
| use ContextFace instead of Qwen2.5-VL-7B (quote) | ContextFace · SFEW-C zero-shot with quote | UAR: 90.20 -> 93.21 | Improved |
| use ContextFace predictions instead of mean expression baseline | ContextFace · SFEW-C expression prediction (situation) | Frechet distance: 14.38 -> 0.81 | Improved |
| use ContextFace instead of mean baseline for quote L2 | ContextFace · SFEW-C expression prediction (quote) | L2 error: 0.12 -> 0.10 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- BLIP-13B — BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation (ICML 2022) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Facial Dynamics in Video: Instruction Tuning for Improved Facial Expression Perception and Contextual Awareness (AAAI 2026 · similarity 0.89)
- DEEPTalk: Dynamic Emotion Embedding for Probabilistic Speech-Driven 3D Face Animation (AAAI 2025 · similarity 0.88)
- Customizing Visual Emotion Evaluation for MLLMs: An Open-vocabulary, Multifaceted, and Scalable Approach (ICLR 2026 · similarity 0.88)
- Facial-R1: Aligning Reasoning and Recognition for Facial Emotion Analysis (AAAI 2026 · similarity 0.87)
- FlowVQTalker: High-Quality Emotional Talking Face Generation through Normalizing Flow and Quantization (CVPR 2024 · similarity 0.86)
- ExpCLIP: Bridging Text and Facial Expressions via Semantic Alignment (AAAI 2024 · similarity 0.86)
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.
