DEL: Dense Event Localization for Multi-modal Audio-Visual Understanding
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915364125802496 |
|---|---|
| author | Ahmadian, Mona Shirian, Amir Guerin, Frank Gilbert, Andrew |
| author_facet | Ahmadian, Mona Shirian, Amir Guerin, Frank Gilbert, Andrew |
| contents | Real-world videos often contain overlapping events and complex temporal dependencies, making multimodal interaction modeling particularly challenging. We introduce DEL, a framework for dense semantic action localization, aiming to accurately detect and classify multiple actions at fine-grained temporal resolutions in long untrimmed videos. DEL consists of two key modules: the alignment of audio and visual features that leverage masked self-attention to enhance intra-mode consistency and a multimodal interaction refinement module that models cross-modal dependencies across multiple scales, enabling high-level semantics and fine-grained details. Our method achieves state-of-the-art performance on multiple real-world Temporal Action Localization (TAL) datasets, UnAV-100, THUMOS14, ActivityNet 1.3, and EPIC-Kitchens-100, surpassing previous approaches with notable average mAP gains of +3.3%, +2.6%, +1.2%, +1.7% (verb), and +1.4% (noun), respectively. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_23196 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DEL: Dense Event Localization for Multi-modal Audio-Visual Understanding Ahmadian, Mona Shirian, Amir Guerin, Frank Gilbert, Andrew Computer Vision and Pattern Recognition Real-world videos often contain overlapping events and complex temporal dependencies, making multimodal interaction modeling particularly challenging. We introduce DEL, a framework for dense semantic action localization, aiming to accurately detect and classify multiple actions at fine-grained temporal resolutions in long untrimmed videos. DEL consists of two key modules: the alignment of audio and visual features that leverage masked self-attention to enhance intra-mode consistency and a multimodal interaction refinement module that models cross-modal dependencies across multiple scales, enabling high-level semantics and fine-grained details. Our method achieves state-of-the-art performance on multiple real-world Temporal Action Localization (TAL) datasets, UnAV-100, THUMOS14, ActivityNet 1.3, and EPIC-Kitchens-100, surpassing previous approaches with notable average mAP gains of +3.3%, +2.6%, +1.2%, +1.7% (verb), and +1.4% (noun), respectively. |
| title | DEL: Dense Event Localization for Multi-modal Audio-Visual Understanding |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2506.23196 |