Offline Signature Verification

Convolutional Siamese Network

Determination of the Precision-Recall point:

• It is an important case to determine that the signature is false in the banking sector.
• Therefore, keeping the "Recall" value high is a priority.
• However, the "Precision" value should be in a range that prevents the customer from signing again and again.

REFERENCES

1- https://arxiv.org/pdf/1707.02131.pdf

2- https://github.com/sounakdey/SigNet (Tensorflow -original)

3- https://github.com/Aftaab99/OfflineSignatureVerification (Pytorch)

4- https://www.kaggle.com/robinreni/signature-verification-dataset

5- https://en.wikipedia.org/wiki/Sensitivity_and_specificity

6- https://machinelearningmastery.com/roc-curves-and-precision-recall-curves-for-classification-in-python/

APPENDIX

• ROC Curves summarize the trade-off between the true positive rate and false positive rate for a predictive model using different probability thresholds.
• Precision-Recall curves summarize the trade-off between the true positive rate and the positive predictive value for a predictive model using different probability thresholds.
• ROC curves are appropriate when the observations are balanced between each class
• Precision-recall curves are appropriate for imbalanced datasets.