Apple Classification using Haralick Features and KNN

Features

• Preprocessing: Includes resize, convert to grayscale, and preparing images for feature extraction.
• Feature Extraction: Utilizes Haralick texture features with d=(1,2,3) and theta=(0,45,90,135) combination.
• Classification: Implements KNN for classifying apple ripeness levels with different K values to each dataset combination.
• Dashboard: Interactive Streamlit dashboard for visualizing results and exploring data.

Installation

1. Clone the repository:

   git clone https://github.com/Bagusdevaa/Apple-Classfication-using-Harralick-and-KNN.git
   cd Apple-Classfication-using-Harralick-and-KNN

2. Set up a Python virtual environment:

   python -m venv skripsi
   skripsi\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

Usage

Running the Dashboard

1. Navigate to the src/dashboard directory.
2. Run the Streamlit app:

   streamlit run app.py

3. Open the provided URL in your browser to access the dashboard.

Notebooks

• Explore the Jupyter notebooks in the notebooks/ directory for preprocessing, feature extraction, and classification.

Dataset

The dataset is organized into folders based on ripeness levels (20%, 40%, 60%, 80%, 100%). Each folder contains images of apples at the corresponding ripeness level.

Results

• The dashboard provides visualizations of classification results, including accuracy, precision, recall, f1-score, confusion matrices, and Nearest Neighbors Visualizatoin.
• The best KNN hyperparameters and K value of KNN are determined through experimentation and displayed in the dashboard.

License

This project is for educational purposes and is not licensed for commercial use.