👤 Gender Classification from Fingerprints

📖 Introduction

This project is a deep learning application that predicts gender from fingerprint images.
It uses a Convolutional Neural Network (CNN) trained on the SOCOFing dataset and provides a Streamlit web interface for uploading and classifying fingerprints in real time.

📑 Table of Contents

• Introduction
• Dataset
• Installation
• Usage
• Features
• Project Structure
• Dependencies
• Configuration
• Examples
• Troubleshooting
• Contributors
• License

📂 Dataset

This project uses the SOCOFing dataset, which contains 6,000 fingerprint images (real and synthetically altered).

Option 1: Manual Download

1. Download the dataset from Kaggle:
   👉 SOCOFing Dataset on Kaggle

2. Extract the dataset into your project folder so the structure looks like this:

   SOCOFing/
   ├── Real/
   └── Altered/
       ├── Altered-Easy/
       ├── Altered-Medium/
       └── Altered-Hard/
   

Option 2: Kaggle API (Automatic Download)

1. Install the Kaggle API:

   pip install kaggle

2. Create a Kaggle API token:

   • Go to your Kaggle Account Settings
   • Click Create New API Token
   • This downloads a file named kaggle.json

3. Place kaggle.json in:

   • Linux/Mac: ~/.kaggle/kaggle.json
   • Windows: C:\Users\<YourUsername>\.kaggle\kaggle.json

4. Download and unzip the dataset automatically:

   kaggle datasets download -d ruizgara/socofing
   unzip socofing.zip -d .

Now your dataset is ready in the SOCOFing/ folder.

⚙️ Installation

1. Clone this repository:

   git clone https://github.com/AryanGupta5084/DL.git
   cd DL

2. Install dependencies:

   pip install -r requirements.txt

3. Ensure you have the SOCOFing dataset placed in the correct folder structure (see above).

🚀 Usage

1. Train the Model

Run the training script to build and save the CNN model (trained_model.h5):

python train.py

2. Run the Web App

Once the model is trained, launch the Streamlit app:

streamlit run app.py

3. Upload Fingerprints

• Upload a fingerprint image (.jpg, .jpeg, .png)
• The app will classify it as Male or Female

✨ Features

• CNN-based gender classification from fingerprints
• Training on multiple difficulty levels of the SOCOFing dataset
• Real-time predictions via a Streamlit web interface
• Automatic preprocessing (grayscale conversion, resizing, normalization)
• Model saved in trained_model.h5 for reuse

📂 Project Structure

DL/
├── app.py              # Streamlit web app for predictions
├── train.py            # CNN training script
├── requirements.txt    # Dependencies
├── trained_model.h5    # Saved model (created after training)
└── SOCOFing/           # Dataset (not included in repo)

📦 Dependencies

Main dependencies (from requirements.txt):

• numpy
• pandas
• matplotlib, seaborn
• scikit-learn
• tensorflow, keras
• opencv-python
• streamlit
• Pillow

Install all with:

pip install -r requirements.txt

⚙️ Configuration

• Image size: 96×96 pixels
• Batch size: 128
• Epochs: 30 (with early stopping)
• Optimizer: Adam (learning rate = 0.001)
• Loss function: Categorical Crossentropy

🖼️ Examples

1. Upload a fingerprint:
   
2. Predicted Output in Streamlit:

   Predicted Gender: Male
   

🛠️ Troubleshooting

• Error: trained_model.h5 not found → Run python train.py before launching the app.
• Dataset missing → Ensure the SOCOFing dataset is properly downloaded and structured.
• OpenCV errors → Check that fingerprint images are valid grayscale images.

👥 Contributors

• Aryan Gupta — Developer & Maintainer