IMAGE FORGERY DETECTION USING DEEP LEARNING

Abstract

Abstract—This study present an innovative approach for Aadhar card authentication and age estimation leveraging deep learning and Convolutional Neural Networks (CNNs). Our methodology integrates two key components: image processing and text extraction. Initially, we preprocess Aadhar card images using grayscale conversion, resizing, and normalization to facilitate effective CNN-based analysis. Subsequently, a pre-trained CNN model is employed for card authentication, enabling the classification of Aadhar cards into valid or invalid categories based on learned features. Additionally, we utilize Optical Character Recognition (OCR) techniques to extract textual information from the cards. Specifically, we focus on extracting the date of birth (DOB) from the textual content. By employing regular expressions, we accurately identify DOB instances within the extracted text. This information is then utilized to estimate the age of the Aadhar card holder. Our experimental results demonstrate the efficacy of the proposed approach in accurately verifying Aadhar cards and estimating the age of the cardholders. The integration of deep learning and OCR technologies showcases a promising avenue for enhancing identification and authentication systems, particularly in scenarios requiring age estimation for regulatory compliance or age-sensitive services.