LongViTU: Instruction Tuning for Long-Form Video Understanding
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866910895846719488 |
|---|---|
| author | Wu, Rujie Ma, Xiaojian Ci, Hai Fan, Yue Wang, Yuxuan Zhao, Haozhe Li, Qing Wang, Yizhou |
| author_facet | Wu, Rujie Ma, Xiaojian Ci, Hai Fan, Yue Wang, Yuxuan Zhao, Haozhe Li, Qing Wang, Yizhou |
| contents | This paper introduces LongViTU, a large-scale (~121k QA pairs, ~900h videos), automatically generated dataset for long-form video understanding. We propose a systematic approach that organizes videos into a hierarchical tree structure for QA generation and incorporates self-revision mechanisms to ensure high-quality QA pairs. Each QA pair in LongViTU features: 1) long-term context (average certificate length of 4.6 minutes); 2) rich knowledge and condensed reasoning (commonsense, causality, planning, etc.)). We also offer explicit timestamp annotations of relevant events for each QA pair. We have conducted extensive human studies on LongViTU, and the results prove the quality of our dataset. To better evaluate the challenges posed by LongViTU's emphasis on long-term context and condensed reasoning, we manually curate a subset of LongViTU into a benchmark. Evaluations using a state-of-the-art open-source model (LongVU), a proprietary model (Gemini-1.5-Pro), and human annotators yield GPT-4 scores of 49.9, 52.3, and 81.0, respectively, underscoring the substantial difficulty presented by LongViTU questions. Performing supervised fine-tuning (SFT) of LongVU and LLaVA-Video on LongViTU data results in average performance gains of 2.5% and 3.7%, respectively, across a suite of long video understanding benchmarks (EgoSchema, VideoMME-Long, MLVU, LVBench). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_05037 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | LongViTU: Instruction Tuning for Long-Form Video Understanding Wu, Rujie Ma, Xiaojian Ci, Hai Fan, Yue Wang, Yuxuan Zhao, Haozhe Li, Qing Wang, Yizhou Computer Vision and Pattern Recognition Machine Learning This paper introduces LongViTU, a large-scale (~121k QA pairs, ~900h videos), automatically generated dataset for long-form video understanding. We propose a systematic approach that organizes videos into a hierarchical tree structure for QA generation and incorporates self-revision mechanisms to ensure high-quality QA pairs. Each QA pair in LongViTU features: 1) long-term context (average certificate length of 4.6 minutes); 2) rich knowledge and condensed reasoning (commonsense, causality, planning, etc.)). We also offer explicit timestamp annotations of relevant events for each QA pair. We have conducted extensive human studies on LongViTU, and the results prove the quality of our dataset. To better evaluate the challenges posed by LongViTU's emphasis on long-term context and condensed reasoning, we manually curate a subset of LongViTU into a benchmark. Evaluations using a state-of-the-art open-source model (LongVU), a proprietary model (Gemini-1.5-Pro), and human annotators yield GPT-4 scores of 49.9, 52.3, and 81.0, respectively, underscoring the substantial difficulty presented by LongViTU questions. Performing supervised fine-tuning (SFT) of LongVU and LLaVA-Video on LongViTU data results in average performance gains of 2.5% and 3.7%, respectively, across a suite of long video understanding benchmarks (EgoSchema, VideoMME-Long, MLVU, LVBench). |
| title | LongViTU: Instruction Tuning for Long-Form Video Understanding |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2501.05037 |