Signature Forgery Detection and Verification using Deep Learning Techniques
Sri Chakradhar Nossam, Rishi Anirudh Katakam, Gopa Pulastya, Sarada Jayan · 2024
Signature Verification plays a vital role in this current digital era. This work is aimed at studying how signatures can be verified using image processing and Deep Learning methods for forgery detection. We implement Sobel edge detection techniques to allocate the features coming from original signature images and then fake ones, followed by the statistical analysis that inquires the color channels, shape and the average pixel values. Besides that, we do the LSTM and CNN models, which are used to recognize if the given signature is real or forged. LSTM model bases on the finding of flattened images, instead the convolutional layers are being applied for the feature extraction in CNN model. In this process, we plan to create a way to examine images and signatures in order to ensure the safety and higher level of security. We have attained a good accuracy by using the Convolution Neural Network.