ML / Done
Tomato Leaf Disease Detection
Overview
Developed a lightweight deep learning model using TensorFlow Lite to detect diseases in tomato leaves from images. The model, based on transfer learning with MobileNetV2, achieves high accuracy with a compact size, suitable for deployment on resource-constrained devices.
Challenge
Optimizing a high-accuracy model to maintain performance within a compact 5MB size, handling a large dataset with 11 classes, and keeping detection steady across varied image conditions such as lighting and angles.
Result
The model achieved 99.4% accuracy in detecting 11 classes of tomato leaf diseases using a dataset of over 17,000 images. Deployed on Streamlit, the application allows users to upload images and receive real-time disease predictions, demonstrating potential for agricultural diagnostic tools.
Key Statistics
99.4%
Model Accuracy
17,000+ images
Dataset Size
5MB
Model Size
11
Number of Classes
Technologies
Machine Learning
-
Python -
TensorFlow -
Keras
Data Processing
-
NumPy -
Pandas
Visualization
-
Seaborn
Deployment
-
Streamlit - TensorFlow Lite
Gallery
Disease Detection Output
Visualization of detected tomato leaf disease with confidence scores.
Streamlit Interface
Interactive Streamlit app showing real-time disease prediction results.
