when u r selecting the fingerprint image from A- and uploading it will obviously give A- only...... what are we supposed to do when professor say detect my blood group? how to obtain his fingerprint image, I'll somehow find an app or smtg to get finger image and then in which cluster should i place his image ??, ntg is making sense here.
The model is trained on a specific set of images labeled with known blood groups (like A-, B+, etc.). The model identifies patterns in these images to classify new ones. However, the model is currently limited to detecting blood groups based on image data from a controlled dataset. This means the model needs to be trained with a substantial dataset of fingerprint images that correspond to verified blood group types to generalize well to unseen fingerprints. The model doesn’t use explicit "clusters" in the manual sense but instead relies on the underlying deep learning model (MobileNetV2) to automatically classify images into predefined blood group categories. If the professor’s fingerprint closely resembles patterns seen in the training data for, say, "A-," the model would likely classify it as such. However, if the model is uncertain due to lack of data or variability in fingerprint patterns, it may misclassify or give a low-confidence prediction. Kindly note that this is only a Final Year Project. Not a real time application. This project demonstrates a proof of concept - it’s not yet a fully validated medical application. To improve reliability in real-world use, future work should involve gathering a large, verified dataset of fingerprint images across different blood groups, allowing the model to generalize better. The technology’s limitations should be clearly communicated in practical applications to manage expectations
Did you get fingerprint images using scanner or something else and then convert it into gery scale.
Referred from Kaggle.
when u r selecting the fingerprint image from A- and uploading it will obviously give A- only...... what are we supposed to do when professor say detect my blood group? how to obtain his fingerprint image, I'll somehow find an app or smtg to get finger image and then in which cluster should i place his image ??, ntg is making sense here.
The model is trained on a specific set of images labeled with known blood groups (like A-, B+, etc.). The model identifies patterns in these images to classify new ones. However, the model is currently limited to detecting blood groups based on image data from a controlled dataset. This means the model needs to be trained with a substantial dataset of fingerprint images that correspond to verified blood group types to generalize well to unseen fingerprints.
The model doesn’t use explicit "clusters" in the manual sense but instead relies on the underlying deep learning model (MobileNetV2) to automatically classify images into predefined blood group categories. If the professor’s fingerprint closely resembles patterns seen in the training data for, say, "A-," the model would likely classify it as such. However, if the model is uncertain due to lack of data or variability in fingerprint patterns, it may misclassify or give a low-confidence prediction.
Kindly note that this is only a Final Year Project. Not a real time application. This project demonstrates a proof of concept - it’s not yet a fully validated medical application.
To improve reliability in real-world use, future work should involve gathering a large, verified dataset of fingerprint images across different blood groups, allowing the model to generalize better.
The technology’s limitations should be clearly communicated in practical applications to manage expectations
How did you collect all dataset of fingerprint ? Can you share with me .
Referred from Kaggle.
cant it detect live?? with a fingerprint sensor
@@pranzz30 Sorry, Not Possible.
What is the cost of this project?
Rs.5000/