FootArch Classification using Vision Transformers & Grey Wolf Optimizer

The goal was to classify foot arch types specifically detecting Pes Planus (flatfoot) versus normal arches, using image-based deep learning features combined with classical machine learning (Random Forest Classifier).

For Feature Extraction: Leveraged pretrained Vision Transformers (ViT) from the timm library to extract high-dimensional feature embeddings from medical foot images.

Optimization: Applied the Grey Wolf Optimizer (GWO) for hyperparameter tuning of a Random Forest Classifier, improving both precision and generalization.

Model Performance:

- Accuracy: 97.04%

- Recall: 93.83%

- Precision: 100.00%

- F1-Score: 96.82%