A deep learning project that uses a Convolutional Neural Network (CNN) to predict a person’s blood group from fingerprint images. Built with PyTorch and Flask, this web app allows users to upload fingerprint images and get instant blood group predictions.
- Uses CNN to analyze fingerprint images.
- Supports 8 blood group classes.
- Includes training, validation, and testing pipeline.
- Web interface for uploading fingerprint images and getting predictions.
- Confusion matrix and performance metrics visualization.
The dataset used for training was obtained from Kaggle: 🔗Fingerprint-Based Blood Group Dataset: https://www.kaggle.com/datasets/rajumavinmar/finger-print-based-blood-group-dataset Please download it separately and place it in the appropriate folder before training.
-
Clone the Repository git clone https://github.com/MreenalRanjan23/Blood_Group_Detection_Using_Fingerprint.git cd Blood-Group-Detection-Using-Fingerprint
-
Install Dependencies pip install -r requirements.txt
-
Start the Flask App python run.py
-
Visit in Browser http://localhost:5000
To train the model run: python Scripts/train_model.py Ensure the dataset is placed correctly before training.
Simple login and registration system using SQLAlchemy and PostgreSQL. Update the SQLALCHEMY_DATABASE_URI in app/init.py with your own credentials.
This project is licensed under the MIT License. See the LICENSE file for more details.