Tomato-Leaf-Diseases-TensorFlow-Object-Detection

This repository contains code for detecting tomato leaf diseases using TensorFlow Object Detection API. The project includes scripts to set up the environment, train a custom object detection model using TensorFlow 2.x, convert the trained model to TensorFlow Lite format, and perform inference using the TFLite model.

Installation

1. Clone the TensorFlow models repository:

   !pip uninstall Cython -y 
   !git clone --depth 1 https://github.com/tensorflow/models

2. Set up dependencies and environment:

   !pip install pyyaml==5.3
   !pip install /content/models/research/
   !pip install tensorflow==2.8.0
   !pip install tensorflow_io==0.23.1
   # Additional setup steps as per your environment requirements

Training

1. Upload your image dataset and prepare training data:

   from google.colab import drive
   drive.mount('/content/gdrive')
   
   !cp /content/gdrive/MyDrive/ColabNotebooks/data.zip /content
   !cp /content/gdrive/MyDrive/ColabNotebooks/images.zip /content
   !unzip -q data.zip -d /content
   !unzip -q images.zip -d /content

2. Customize the training configuration:

   # Modify model configurations, batch size, and other parameters

3. Train the custom TensorFlow Lite detection model:

   # Run training and save the model in TensorFlow Lite format

Test

1. Test the TensorFlow Lite model:

   # Perform inference with the TFLite model on test images

Example Results from Testing

mean Average Precision

1. Calculate mAP (mean Average Precision):

   # Run mAP calculation on detection results