This project introduces a non-invasive approach to blood group prediction using fingerprint image processing and machine learning. By leveraging Convolutional Neural Networks (CNNs), it classifies fingerprint patterns into eight common blood groups (A+, A-, B+, B-, AB+, AB-, O+, O-), offering a quick and accessible alternative to traditional methods.
-Rapid Blood Group Identification – Provides a fast and accurate alternative to traditional methods.
-Accessibility in Remote Areas – Enables blood group prediction without lab facilities or skilled personnel.
-Integration with Portable Devices – Supports point-of-care diagnostics in clinics and mobile units.
-Safety and Scalability – Reduces contamination risks and ensures adaptability across healthcare settings.
-Biometric and Medical Synergy – Combines biometrics and machine learning for improved diagnostics.
- HTML, CSS, JavaScript
- Flask, SQLAlchemy, SQLite
- TensorFlow/Keras, Google Colab
| Model | Testing Accuracy | Validation Accuracy |
|---|---|---|
| VGG16 | 88.72% | 89.50% |
| AlexNet | 12.47% | 12.49% |
| ResNet50 | 61.19% | 62.70% |
| Hybrid Model (EfficientNetB0 + SVM) | 22.29% | 22.81% |
- Expand the dataset for better generalization.
- Experiment with other models for further accuracy improvements.
- Deploy model in live environment
Email: aamaske50@gmail.com
LinkedIn