Comparison of back propagation neural network and multi-layer perceptron neural network for handwritten signature verification based on accuracy

The aim of this proposed work is to evaluate the performance of the novel back propagation neural network algorithm in predicting a signature image to get a clear form of image and to reduce the forgery by comp aring it with a multi-layer perceptron neural network algorithm. A total number of 100 samples of old signature images which are collected from various classes and labels. These samples are divided into training dataset and test dataset and the accuracy an d skewness values were calculated to quantify the performance of the novel back propagation neural network algorithm and multi-layer perceptron neural network algorithm. Novel back propagation neural network algorithm achieved the accuracy specifically 95. 3% respectively compared to 84.8% of the multi-layer perceptron neural network algorithm of accuracy and skewness. The significance value of 2-tailed is .000 and p less than 0.05. The G power is taken as 0.8. It is observed that the novel back propagation neural network algorithm performed significantly better than the multi-layer perceptron neural network algorithm in the verification of the signature in the terms on the basis of accuracy and skewness.