A CNN-Based Approach for Image Recognition and Categorizing Facial Skin Conditions

This repository contains a Convolutional Neural Network (CNN) model for classifying facial images into five skin condition categories:

• Acne
• Hyperpigmentation/Discoloration
• Oily Skin
• Dry Skin
• Normal Skin

The model utilizes a Gradio interface to provide a user-friendly platform for predictions, enabling users to upload a facial image and receive the skin condition classification.

---

Features

• CNN Architecture: Designed and trained for accurate classification.
• Gradio Interface: Intuitive web-based UI for uploading images and receiving predictions.
• Custom Dataset: Images collected and labeled for each skin condition.
• Python Implementation: Easy-to-follow, modularized codebase for training, testing, and deploying the model.

---

Installation

1. Clone the Repository

   git clone https://github.com/ruru-lyy/skin-classification-cnn.git  
   cd skin-classification-cnn  

2. Install Dependencies
   Ensure you have Python 3.8 or higher installed. Install the required packages using:

   pip install -r requirements.txt  

3. Download Dataset
   Collect and organize facial images into the data/ directory with the following structure:

   data/  
   ├── sample_acne/  
   ├── sample_hyperpigmentation/  
   ├── sample_oily/  
   ├── sample_dry/  
   └── sample_normal/  
   

4. Train the Model (Optional)
   To train the CNN on your dataset, run:

   python train.py  

   The trained model will be saved in the models/ directory.

---

Model Architecture

The CNN model uses the following components:

• Convolutional layers with ReLU activations.
• Max-pooling layers for dimensionality reduction.
• Dropout for regularization.
• Fully connected layers for classification.

For more details, refer to main.py.

---

Usage

1. Run the Gradio Interface

Start the Gradio app for predictions:

python app.py  

The interface will open in your default web browser or display a local URL in the terminal.

2. Upload Image

Use the Gradio UI to upload an image of a face. The model will predict the skin condition and display the result.

---

App Interface

---

Future Enhancements

• Increase the dataset size for improved accuracy.
• Integrate additional features such as skin-care recommendations based on predictions.
• Deploy the model on the cloud for global access.

---

Contributing

Contributions are welcome! Please create a pull request or open an issue if you encounter any bugs or have feature suggestions.

---

License

This project is licensed under the MIT License. See the LICENSE file for more details.

---

Feel free to use and adapt this project. If you find it helpful, please give this repository a ⭐ on GitHub! 😊