Multimodal Alignment with Cross-Attentive GRUs for Fine-Grained Video Understanding
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911039073812480 |
|---|---|
| author | Kim, Namho Kim, Junhwa |
| author_facet | Kim, Namho Kim, Junhwa |
| contents | Fine-grained video classification requires understanding complex spatio-temporal and semantic cues that often exceed the capacity of a single modality. In this paper, we propose a multimodal framework that fuses video, image, and text representations using GRU-based sequence encoders and cross-modal attention mechanisms. The model is trained using a combination of classification or regression loss, depending on the task, and is further regularized through feature-level augmentation and autoencoding techniques. To evaluate the generality of our framework, we conduct experiments on two challenging benchmarks: the DVD dataset for real-world violence detection and the Aff-Wild2 dataset for valence-arousal estimation. Our results demonstrate that the proposed fusion strategy significantly outperforms unimodal baselines, with cross-attention and feature augmentation contributing notably to robustness and performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_03531 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Multimodal Alignment with Cross-Attentive GRUs for Fine-Grained Video Understanding Kim, Namho Kim, Junhwa Computer Vision and Pattern Recognition Artificial Intelligence Fine-grained video classification requires understanding complex spatio-temporal and semantic cues that often exceed the capacity of a single modality. In this paper, we propose a multimodal framework that fuses video, image, and text representations using GRU-based sequence encoders and cross-modal attention mechanisms. The model is trained using a combination of classification or regression loss, depending on the task, and is further regularized through feature-level augmentation and autoencoding techniques. To evaluate the generality of our framework, we conduct experiments on two challenging benchmarks: the DVD dataset for real-world violence detection and the Aff-Wild2 dataset for valence-arousal estimation. Our results demonstrate that the proposed fusion strategy significantly outperforms unimodal baselines, with cross-attention and feature augmentation contributing notably to robustness and performance. |
| title | Multimodal Alignment with Cross-Attentive GRUs for Fine-Grained Video Understanding |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2507.03531 |