This project demonstrates an image classification system using a Support Vector Machine (SVM) model trained on a custom dataset with three categories: Rugby Ball Leather, Ice Cream Cone, and Sunflower.
- Preprocesses and resizes input images to 150x150
- Trains an SVM model with hyperparameter tuning using GridSearchCV
- Achieves ~91% accuracy on test data
- Deploys a user-friendly Streamlit app to upload and classify new images
- Displays predicted class and probability distribution