Some biometric cryptographic applications such as fuzzy identity-based identification (FIBI) and fuzzy commitment require ordered and fixed-length bit-string-like IrisCode as input. However, fingerprint minutiae representation (e.g. ISO minutiae format) is unordered and variable in size. Such a characteristic is inapplicable to the aforementioned applications. One of the feasible solutions is to convert minutiae into ordered and fixed-length bit-string, namely point-to-string conversion. The point-to-string conversion has attracted much attention and a number of proposals have been reported in literature over the past decade. Furthermore, the topic of point-to-string conversion continues to gain the interest from the research community lately. In this chapter, the point-to-string conversion methods proposed in early stage are revisited to be served as a background study. Thereafter, a review of recent development on point-to-string conversion is presented. More specifically, two recently proposed methods (i.e. Kernel-Learning and Bag-of-Minutiae) are introduced in detail. Finally, conclusion is given to summarize the challenges, and future prospect in this research topic.
|Title of host publication||Bio-Inspired Computing Models and Algorithms|
|Publisher||World Scientific Publishing Co.|
|Number of pages||31|
|Publication status||Published - 2019 Jan 1|
Bibliographical notePublisher Copyright:
© 2019 by World Scientific Publishing Co. Pte. Ltd.
All Science Journal Classification (ASJC) codes
- Computer Science(all)