Rebuttal to “Comments on Random Distance Method Generating Unimodal and Multimodal Cancelable Biometric Features”

Abstract

This rebuttal highlights major flaws in the comments made by Lamba due to possible oversight in understanding the scheme given by Kaur and Khanna. Firstly, the random distance method (RDM) does not suffer from any mathematical fallacy. It is a unique technique to generate cancelable biometric templates. Secondly, the size of the original log-Gabor feature vector is reduced by half due to the inherent nature of the RDM scheme, and it does not affect the inter-and intra user variations in any adverse way. An experimental analysis is also performed to showcase that RDM maintains inter and intra-user variations. Thirdly, the comment correspondence author has not noticed that the sole purpose of using the random grid (RG) with OR operation is to increase the entropy of the log-Gabor features, which have a low dynamic range. The issue of partial information revelation is out of context here as salting is not used for hiding information. Finally, the author of comment correspondence has failed to understand that padding is used in the implementation of median filters, which is very common in the filtering process.