A graphic-based cryptographic model for authentication

Abstract

Conventional password has been used for authentication for a long time due to its advantages. However, it drawbacks such as weak or unrecalled passwords has frequently compromised security. This paper presents a graphic-based cryptographic model (GBCM) with enhanced usability and security. The GBCM model consists of registration and verification stages that users must successfully complete in order to be authenticated. A hybrid technique based on recognition, cued and pure recall was adopted. The GBCM security is enhanced by using a three-level authentication mode, secret key, operator and scrambling of grid cells, thereby mitigating shoulder surfing attack; usability is enhanced using grid cell identity (ID) and images. Consequently, the implementation shows that, out of 18 users registration done with the system, 83.33% remembered their images, 83.33% also remembered their secret keys while 88.89% remembered their operator, resulting in 77.78% successful login. The login success proves user friendliness of the GBCM system.