Learning Cross-Joint Attention for Generalizable Video-Based Seizure Detection
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915889255809024 |
|---|---|
| author | Zamzam, Omar Medani, Takfarinas Chinara, Chinmay Leahy, Richard |
| author_facet | Zamzam, Omar Medani, Takfarinas Chinara, Chinmay Leahy, Richard |
| contents | Automated seizure detection from long-term clinical videos can substantially reduce manual review time and enable real-time monitoring. However, existing video-based methods often struggle to generalize to unseen subjects due to background bias and reliance on subject-specific appearance cues. We propose a joint-centric attention model that focuses exclusively on body dynamics to improve cross-subject generalization. For each video segment, body joints are detected and joint-centered clips are extracted, suppressing background context. These joint-centered clips are tokenized using a Video Vision Transformer (ViViT), and cross-joint attention is learned to model spatial and temporal interactions between body parts, capturing coordinated movement patterns characteristic of seizure semiology. Extensive cross-subject experiments show that the proposed method consistently outperforms state-of-the-art CNN-, graph-, and transformer-based approaches on unseen subjects. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_23757 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Learning Cross-Joint Attention for Generalizable Video-Based Seizure Detection Zamzam, Omar Medani, Takfarinas Chinara, Chinmay Leahy, Richard Computer Vision and Pattern Recognition Automated seizure detection from long-term clinical videos can substantially reduce manual review time and enable real-time monitoring. However, existing video-based methods often struggle to generalize to unseen subjects due to background bias and reliance on subject-specific appearance cues. We propose a joint-centric attention model that focuses exclusively on body dynamics to improve cross-subject generalization. For each video segment, body joints are detected and joint-centered clips are extracted, suppressing background context. These joint-centered clips are tokenized using a Video Vision Transformer (ViViT), and cross-joint attention is learned to model spatial and temporal interactions between body parts, capturing coordinated movement patterns characteristic of seizure semiology. Extensive cross-subject experiments show that the proposed method consistently outperforms state-of-the-art CNN-, graph-, and transformer-based approaches on unseen subjects. |
| title | Learning Cross-Joint Attention for Generalizable Video-Based Seizure Detection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.23757 |