Salvato in:
| Autori principali: | , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2512.03405 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908907959484416 |
|---|---|
| author | Wu, Jiangtao Li, Shihao Bian, Zhaozhou Chen, Jialu Wen, Runzhe Ping, An He, Yiwen Wang, Jiakai Zhang, Yuanxing Liu, Jiaheng |
| author_facet | Wu, Jiangtao Li, Shihao Bian, Zhaozhou Chen, Jialu Wen, Runzhe Ping, An He, Yiwen Wang, Jiakai Zhang, Yuanxing Liu, Jiaheng |
| contents | Understanding visual differences between dynamic scenes requires the comparative perception of compositional, spatial, and temporal changes--a capability that remains underexplored in existing vision-language systems. While prior work on Image Difference Captioning (IDC) has enabled models to describe semantic changes between static images, these approaches fail to capture motion continuity, event evolution, or editing consistency over time. We introduce the ViDiC (Video Difference Captioning) task and its corresponding ViDiC-1K dataset, designed to evaluate the ability of Multimodal Large Language Models (MLLMs) to provide fine-grained descriptions of similarities and differences between video pairs. ViDiC-1K comprises 1,000 curated video pairs annotated with over 4,000 comparative checklist items, covering seven categories: subject, style, background, cinematography, motion, location, and playback techniques. To ensure reliable evaluation, we propose a dual-checklist framework that measures the accuracy of similarity and difference separately, based on the LLM-as-a-Judge protocol. Experiments on nineteen representative multimodal models reveal a significant performance gap in their comparative description and difference perception abilities. We hope ViDiC-1K can be a challenging benchmark that lays a solid foundation for advancing video understanding, edit awareness, and comparative reasoning in multimodal intelligence. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_03405 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | ViDiC: Video Difference Captioning Wu, Jiangtao Li, Shihao Bian, Zhaozhou Chen, Jialu Wen, Runzhe Ping, An He, Yiwen Wang, Jiakai Zhang, Yuanxing Liu, Jiaheng Computer Vision and Pattern Recognition Understanding visual differences between dynamic scenes requires the comparative perception of compositional, spatial, and temporal changes--a capability that remains underexplored in existing vision-language systems. While prior work on Image Difference Captioning (IDC) has enabled models to describe semantic changes between static images, these approaches fail to capture motion continuity, event evolution, or editing consistency over time. We introduce the ViDiC (Video Difference Captioning) task and its corresponding ViDiC-1K dataset, designed to evaluate the ability of Multimodal Large Language Models (MLLMs) to provide fine-grained descriptions of similarities and differences between video pairs. ViDiC-1K comprises 1,000 curated video pairs annotated with over 4,000 comparative checklist items, covering seven categories: subject, style, background, cinematography, motion, location, and playback techniques. To ensure reliable evaluation, we propose a dual-checklist framework that measures the accuracy of similarity and difference separately, based on the LLM-as-a-Judge protocol. Experiments on nineteen representative multimodal models reveal a significant performance gap in their comparative description and difference perception abilities. We hope ViDiC-1K can be a challenging benchmark that lays a solid foundation for advancing video understanding, edit awareness, and comparative reasoning in multimodal intelligence. |
| title | ViDiC: Video Difference Captioning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.03405 |