Offline signature verification using long short-term memory and histogram orientation gradient

The signing process is a critical step that organizations take to ensure the confidentiality of their data and safeguard it against unauthorized penetration or access. Within last decade, offline handwritten signature research has grown in popularity as common method for human authentication via biometric features. It not an easy task, despite importance this method; struggle such system stem from inability any individual sign same each every time. Additionally, we are indeed interested dataset’s features could affect model's performance; thus, extracted images using histogram orientation gradient (HOG) technique. In paper, suggested long short-term memory (LSTM) neural network model verification, with input USTig CEDAR datasets. Our model’s predictive ability quite outstanding: classification accuracy efficiency LSTM was 92.4% run-time 1.67 seconds 87.7% 2.98 seconds. proposed outperforms other verification approaches K-nearest neighbour (KNN), support vector machine (SVM), convolution (CNN), speeded-up robust (SURF), Harris terms accuracy.