MME-CoF-Pro: Evaluating Reasoning Coherence in Video Generative Models with Text and Visual Hints
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917354964779008 |
|---|---|
| author | Qi, Yu Xu, Xinyi Guo, Ziyu Ma, Siyuan Zhang, Renrui Chen, Xinyan An, Ruichuan Xing, Ruofan Zhang, Jiayi Huang, Haojie Heng, Pheng-Ann Tremblay, Jonathan Wong, Lawson L. S. |
| author_facet | Qi, Yu Xu, Xinyi Guo, Ziyu Ma, Siyuan Zhang, Renrui Chen, Xinyan An, Ruichuan Xing, Ruofan Zhang, Jiayi Huang, Haojie Heng, Pheng-Ann Tremblay, Jonathan Wong, Lawson L. S. |
| contents | Video generative models show emerging reasoning behaviors. It is essential to ensure that generated events remain causally consistent across frames for reliable deployment, a property we define as reasoning coherence. To bridge the gap in literature for missing reasoning coherence evaluation, we propose MME-CoF-Pro, a comprehensive video reasoning benchmark to assess reasoning coherence in video models. Specifically, MME-CoF-Pro contains 303 samples across 16 categories, ranging from visual logical to scientific reasoning. It introduces Reasoning Score as evaluation metric for assessing process-level necessary intermediate reasoning steps, and includes three evaluation settings, (a) no hint (b) text hint and (c) visual hint, enabling a controlled investigation into the underlying mechanisms of reasoning hint guidance. Evaluation results in 7 open and closed-source video models reveals insights including: (1) Video generative models exhibit weak reasoning coherence, decoupled from generation quality. (2) Text hints boost apparent correctness but often cause inconsistency and hallucinated reasoning (3) Visual hints benefit structured perceptual tasks but struggle with fine-grained perception. Website: https://video-reasoning-coherence.github.io/ |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_20194 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MME-CoF-Pro: Evaluating Reasoning Coherence in Video Generative Models with Text and Visual Hints Qi, Yu Xu, Xinyi Guo, Ziyu Ma, Siyuan Zhang, Renrui Chen, Xinyan An, Ruichuan Xing, Ruofan Zhang, Jiayi Huang, Haojie Heng, Pheng-Ann Tremblay, Jonathan Wong, Lawson L. S. Computer Vision and Pattern Recognition Video generative models show emerging reasoning behaviors. It is essential to ensure that generated events remain causally consistent across frames for reliable deployment, a property we define as reasoning coherence. To bridge the gap in literature for missing reasoning coherence evaluation, we propose MME-CoF-Pro, a comprehensive video reasoning benchmark to assess reasoning coherence in video models. Specifically, MME-CoF-Pro contains 303 samples across 16 categories, ranging from visual logical to scientific reasoning. It introduces Reasoning Score as evaluation metric for assessing process-level necessary intermediate reasoning steps, and includes three evaluation settings, (a) no hint (b) text hint and (c) visual hint, enabling a controlled investigation into the underlying mechanisms of reasoning hint guidance. Evaluation results in 7 open and closed-source video models reveals insights including: (1) Video generative models exhibit weak reasoning coherence, decoupled from generation quality. (2) Text hints boost apparent correctness but often cause inconsistency and hallucinated reasoning (3) Visual hints benefit structured perceptual tasks but struggle with fine-grained perception. Website: https://video-reasoning-coherence.github.io/ |
| title | MME-CoF-Pro: Evaluating Reasoning Coherence in Video Generative Models with Text and Visual Hints |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.20194 |