TomatoMato xd Note: This repository represents a learning and experimentation project. It was built to understand machine learning workflows, dataset limitations, and deployment, and is not intended for real-world agricultural use.

Overview: TomatoMan is a small machine learning project created to explore end-to-end ML system design using a mix of image-based and tabular models. The project focuses on understanding: 1)how CNNs behave on real datasets 2)how different models can be wired together 3)how dataset choice affects what a system can meaningfully claim

What This Project Does:

1)Image-based disease classification-- CNN trained on the PlantVillage dataset Supports tomato, potato, and pepper leaf images Experiments with both custom CNNs and pretrained architectures

2)Tabular model (exploratory)-- Random Forest trained on publicly available soil and climate data Used as an auxiliary signal to experiment with tabular ML Included to explore multi-model integration, not as a definitive predictor

3)Fusion logic (experimental)-- Simple logic that combines outputs from the CNN and the Random Forest Implemented to understand decision-level model fusion Results are illustrative and part of the learning process

4)Streamlit application-- Allows uploading a leaf image and entering environment parameters Displays model predictions and confidence scores Demonstrates ML inference and deployment workflow