This project is a deep learning-based web application built using Streamlit that classifies the type and variation of agricultural produce, specifically corn and ginger, using trained ResNet50 models.
- Detect type of produce: Corn or Ginger
- Classify variations such as:
- Corn: Husked, Unhusked, Kernels
- Ginger: Whole, Sliced, In-Context
- Upload an image and get instant predictions
- Powered by fine-tuned ResNet50 models
- Simple browser-based UI with Streamlit
- Architecture: ResNet50 via Transfer Learning
- Framework: TensorFlow / Keras
- Produce Type Model:
produce_type_resnet50.h5 - Variation Model:
produce_variation_resnet.h5 - Input Size: 224x224 RGB images
- Trained on: 6000+ images across both crop categories
Requires Python 3.8 or newer.
git clone https://github.com/<your-username>/PatternRec_Project_Group8.git
cd PatternRec_Project_Group8/streamlit_app
pip install streamlit tensorflow numpy pillow matplotlib
streamlit run app.pyCV Engineer: Jacky He & Muhammad Waseem