sumedhakini/predicting_blood_group_using_fingerprints

🧠 Prediction of Blood Group Using Fingerprints

This web application uses deep learning models (VGG16 & MobileNetV2) to predict a person's blood group based on a fingerprint image.

Built using Gradio, the app provides an interactive and user-friendly interface for quick predictions.

🚀 Features

• —🔍 Predicts one of 8 blood groups: A+, A-, B+, B-, AB+, AB-, O+, O-
• —📊 Uses two different models to ensure accuracy and compare predictions
• —💡 Shows confidence scores and agreement status between models
• —🌐 Fully accessible through Hugging Face Spaces — no installation required
• —🖼 Includes a sample dataset to try the app easily

🧠 Models Used

Both models were trained using TensorFlow and saved as .h5 files. Each model independently predicts the blood group based on the fingerprint image.

🗂 Sample Dataset

A folder named sample_dataset is included in this Space, containing example fingerprint images for each blood group.

Each subfolder corresponds to a blood type:

sample_dataset/ ├── A+/ ├── A-/ ├── B+/ ├── B-/ ├── AB+/ ├── AB-/ ├── O+/ └── O-/

You can try out the app by uploading any image from this dataset!

🖼 Sample Usage

1. 1.Upload a fingerprint image (JPG/PNG)
2. 2.The app preprocesses the image automatically
3. 3.Both models make predictions
4. 4.You receive:
5. 5.Predicted blood group from both models
6. 6.Their confidence scores
7. 7.Whether the models agree or disagree

📁 File Structure

predicting-blood-group-using-fingerprints │ ├── app.py # Gradio app logic ├── VGG16.h5 # Trained VGG16 model ├── MobileNetV2.h5 # Trained MobileNetV2 model ├── requirements.txt # Project dependencies ├── sample_dataset/ # Example fingerprint images │ └── [A+, A-, ..., O-]/ # One folder per blood group ├── README.md # Project overview (this file) └── .gitattributes # Git LFS tracking (auto-created)

🛠 Run Locally (Optional)

To run this app on your own machine:

1. Clone the repo:

git clone https://huggingface.co/spaces/your-username/predicting-blood-group-using-fingerprints
cd predicting-blood-group-using-fingerprints

### 2. Install Dependencies:
pip install -r requirements.txt

### 3. Launch App:
python app.py

Then visit: http://127.0.0.1:7860

---

Feel free to customize the README as needed.