Leaf Disease Classification using Deep Learning Models

I am looking for an experienced machine learning and deep learning expert to build a Leaf Disease Classification Model that can accurately identify and classify plant leaf diseases from images.

Dataset Details:

The dataset contains 87,000 images categorized into 38 different classes, such as Apple healthy, Apple scab, Apple black rot, Tomato botcera, etc.

The dataset is divided into a training and validation set in an 80:20 ratio.


Project Goal:

The objective is to develop a model that can:

1. Identify the correct class (out of 38) to which the leaf belongs.


2. Classify the leaf's health condition as one of the following:

Healthy

Initial stage of disease

Middle stage of disease

Severe damage



3. Analyze the percentage of leaf area affected for severity detection.


4. Provide key performance metrics such as:

Accuracy

Precision

Recall

F1 Score



5. Handle overfitting and underfitting issues explicitly and show how they impact model performance.



Technologies to Use:

Keras API

TensorFlow

Jupyter Notebook


Models to Implement:

1. SVM with HOG (Histogram of Oriented Gradients) for feature extraction


2. SVM with VGG16 (Pre-trained Model) for feature extraction


3. CNN with multiple convolutional layers to directly extract features



Workflow Pipeline:

1. Model Building


2. Training


3. Evaluation


4. Saving the Model


5. Testing and Performance Analysis (Accuracy, Precision, Recall, F1 Score)



Deliverables:

Complete code with proper documentation

Visualization graphs for accuracy and loss

Performance metrics report

Model file (.h5 or .pkl)


Requirements:

Experience in Deep Learning and Computer Vision

Strong understanding of CNN, SVM, and Transfer Learning

Proficiency in TensorFlow and Keras

Ability to handle large datasets