A Framework to Authenticate Signature using Machine Learning Technique

The main objective of this research study is to implement signature verification using machine learning and also to verify the validity of signatures. This study has collected some original signatures and are verified with a random signature. In the proposed model, each signature will have a unique ID. The signatures can be uploaded in different formats like .PNG, .JPG, etc., and the proposed system will check whether the given signature matches with original signature. Initially, pre-processing will be done on the input image to convert it into black and white format and then the x and y coordinates and curves will be extracted to verify the signature. Features like ratio, centroid, eccentricity, solidity, skewness, and kurtosis will be extracted from pre-processed image to characterize various aspects like shape, size, and texture. These features will then help in classifying the genuine and forged signatures. Whenever the input image is given, the original data will be verified and classified either as real signature or forged.