Deep Learning fingerprint recognition using Tensorflow2.
Fingerprint image size is 160x160(500DPI).
Python : 3.7
Tensorflow : >= 2.0
https://github.com/kairess/fingerprint_recognition
You can get sample dataset from here.
Fingerprint Dataset for FVC2000_DB4_B in Kaggle
This dataset was created from FVC2000_DB4_B.
| Release Date | Model Version | Training Dataset | Image Information | Validation accuracy | URL |
|---|---|---|---|---|---|
| 2020-02-21 | v0.1 | 17,859 Images (343 Fingers) | 160 x 160 (500DPI) | 0.9702 | model.7z |
| 2020-04-13 | v0.2_Beta1 | 5,800 Images (203 Fingers) | 160 x 160 (500DPI) | 0.9748 | model_0_2_b1.7z |
Before you train model, you must do preprocess.
Preprocess can generate numpy data from original image.
Please see dataset/README.md file.
You can train your model using source/02_train source.
You can evaluate your model using source/03_evaluation source.
Please donate fingerprint image dataset for research purpose.
Thanks.