Signature Authentication Using AI -ML

Abstract

This paper aims to provide an overview of a fundamental approach to signature detection and identification using standard Python and machine learning libraries. To address the challenge of image recognition, a binary classification process has been used to forecast text or signatures, and signature classifications have been conducted to identify the signer of each signature. The proposed algorithm functions to extract the signature area from scanned documents containing signatures and send it to the trained signature images during the pre-processing stage. The research findings are shown for documents of the same kind, with the signature at the same location. Tensor-slicing techniques applied to NumPy arrays are utilized to pinpoint a particular element within the document. Text and signature-containing areas are extracted using OpenCV tools. Promising results have been observed when certain writers' signatures are recognized. The proposed method illustrates how certain real-world problems can be resolved using classical Python and machine learning libraries.