Explainable Gait Abnormality Detection Using Dual-Dataset CNN-LSTM Models
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914048650510336 |
|---|---|
| author | Agarwal, Parth Chatterjee, Sangaa Kabir, Md Faisal Saha, Suman |
| author_facet | Agarwal, Parth Chatterjee, Sangaa Kabir, Md Faisal Saha, Suman |
| contents | Gait is a key indicator in diagnosing movement disorders, but most models lack interpretability and rely on single datasets. We propose a dual-branch CNN-LSTM framework a 1D branch on joint-based features from GAVD and a 3D branch on silhouettes from OU-MVLP. Interpretability is provided by SHAP (temporal attributions) and Grad-CAM (spatial localization).On held-out sets, the system achieves 98.6% accuracy with strong recall and F1. This approach advances explainable gait analysis across both clinical and biometric domains. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_16472 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Explainable Gait Abnormality Detection Using Dual-Dataset CNN-LSTM Models Agarwal, Parth Chatterjee, Sangaa Kabir, Md Faisal Saha, Suman Computer Vision and Pattern Recognition Gait is a key indicator in diagnosing movement disorders, but most models lack interpretability and rely on single datasets. We propose a dual-branch CNN-LSTM framework a 1D branch on joint-based features from GAVD and a 3D branch on silhouettes from OU-MVLP. Interpretability is provided by SHAP (temporal attributions) and Grad-CAM (spatial localization).On held-out sets, the system achieves 98.6% accuracy with strong recall and F1. This approach advances explainable gait analysis across both clinical and biometric domains. |
| title | Explainable Gait Abnormality Detection Using Dual-Dataset CNN-LSTM Models |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2509.16472 |