This project provides a comprehensive solution for image processing, smart cropping, image embeddings, feature extraction, dataset cleaning, and model training and inference using PyTorch. Additionally, it includes a Gradio-based web application for easy interaction with the trained model.
- Conda Environment: Manage dependencies and environment configuration with Conda.
- Image Processing: Load and filter images efficiently.
- Smart Cropping: Automatically detect and crop images using SIFT key points.
- Image Embeddings: Extract image embeddings using Img2Vec.
- Feature Extraction: Perform PCA for dimensionality reduction and K-means clustering.
- Dataset Manual Cleaning: Tools for cleaning and organizing datasets.
- Torch Model Training: Train custom models using PyTorch.
- Model Inference: Make predictions with trained models.
- Web App: Interact with the model via a Gradio-based web application.