Neural integrated Markov model for effective script identification and classification in biometric system

Abstract

Biometric scripts are required to be adaptable and robust for differentiating human input for preventing forging. However, based on the individual biometrics system's functionality, the script identification and classification are vital for input detection. Therefore, for improving the adaptability and improved analysis, this article introduces an Integrated Classification model (ICM) using Neural Markov Algorithm (NMA) for biometric systems. The proposed model jointly performs script detection and classification for different biometric functions. The flaws in the process are identified using output validation and the error between output and input. Therefore, this random change in error is computed using the Markov model for detection based on which the classification as robust or non-robust is decided. For improving the robust type classification, the current error level is accounted for and script flaws and improvements are recommended. Based on the script improvement further testing is performed for the detection phase using Markov stochastic computation. This detection is alone performed for preventing non-robust classifications for varying inputs. Therefore, the maximum robust combinations are identified across different biometric scripts that suit the design purpose.