Abstract

This thesis describes a web-based application that will enable users to access their networked home appliances over the Internet in an easy, secured, accessible and cost effective manner, using the user's iris image only for authentication. As Internet is increasingly gaining significance and popularity in our daily lives, various home networking technologies also started gaining importance from consumers, which helped in facilitating interoperability, sharing of services and exchange of information between different electronic devices at home. As a result, the demand to be able to access home appliances or security cameras over the Internet gradually grew. In this research, we propose an efficient, secured, low-cost and user-friendly method to access networked home appliances over the Internet, providing strong, well integrated, three levels of security to the whole application and user data. According to our design, the user's iris data after hashing (using BioHashing) is sent through a secure communication channel utilizing Secure Sockets Layer v-3.0. The deterministic feature sequence from the iris image is extracted using 1D log-Gabor filters and while performing BioHashing, the orthonormalization of the pseudorandom number is implemented employing Gram-Schmidt orthonormalization algorithm. In addition to this protected data transfer mechanism, we propose the design of an Authentication Server that can be shared among multiple homes, allowing numerous users to access their home appliances in a trouble-free and secured manner. It can also bring down the cost of commercial realization of this endeavor and increase its accessibility without compromising on system security. We demonstrate that the recognition efficiency of this system is computationally effective with equal error rate (EER) of 0% and 6.75% (average) in two separate conditions on CASIA 1 and CASIA 2 iris image datasets.