Fingerprint-Spoof-Detector-based-on-LBP

Implementation:

1. Conversion of each image into grayscale before we extract the LBP features.
2. Extraction of LBP features from the LocalBinaryPattern implementation found in scikit-image.
3. SVC is used as it tries to classify the classes based on maximum margin by taking extreme points.
4. Performed GridSearch on SVC to find out that non-linear kernel -RBF perform well when compared to the linear kernel.
5. Best parameters fitted to our model.
6. We can see the result our model based on our selected performance metrics.

Performance Metrics:

1. Accuracy
2. Precision
3. Recall
4. Confusion matrix

Programming/Libraries:

1. Python
2. opencv
3. sklearn

References:

1. pyimagesearch
2. scikit-learn