Emo-CFG: Video Emotion Recognition Dataset for Foundation Models

Emo-CFG is a dataset for emotion-centric video foundation models, accepted at the NeurIPS 2025 conference. It was created by researchers from Nankai University, Pengcheng Laboratory, and Kuaishou Technology. The dataset was last updated on December 7, 2025.

Use Cases

• Train video foundation models for emotion recognition based on the dataset's affective reasoning framework.
• Benchmark multimodal reasoning systems on emotion-centric tasks based on the described 'Affective-Tree Reasoning'.
• Develop emotion-aware AI applications for video content based on the dataset's focus on emotion-centric models.

Strengths

• Dataset is associated with a NeurIPS 2025 conference acceptance, suggesting peer-reviewed quality.
• Created by a collaboration between academic (Nankai University) and industry (Kuaishou Technology) researchers.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

KlingTeam (affiliated with KlingAI, Kuaishou Technology, Nankai University, Pengcheng Laboratory)

Freshness

Last updated 2025-12-07 00:05:11; freshness should be verified.