Age Classification through Handwriting Analysis using Convolutional Neural Networks and Pre-Segmented Offline Handwritten Gurumukhi Characters

Handwriting analysis is increasingly gaining prominence across various domains, including forensics, education, and biometrics, owing to its invaluable insights. This article introduces an innovative approach to age classification based on offline handwriting in Gurumukhi script, employing a Convolutional Neural Network (CNN) architecture, specifically the Inception-V3 model. The method utilizes pre-segmented offline Gurumukhi characters, commonly employed in the Punjabi language, as input features. This pre-segmentation process ensures precise and efficient extraction of individual characters from handwriting samples. The CNN model undergoes training on a substantial dataset of annotated Gurumukhi characters, encompassing a diverse range of age groups. The dataset comprises 74,760 Gurumukhi characters contributed by 267 writers, with 137 males and 130 females, each providing 280 characters. By inputting the extracted features into the CNN model, our approach accurately predicts individuals’ age groups based on their handwriting samples. Experimental results underscore the efficacy of the proposed method, achieving an impressive accuracy rate of 92.7% in age classification. This high classification accuracy unequivocally validates the potential of our method for age classification in handwriting analysis. It underscores its applicability in forensic investigations and document verification scenarios, highlighting the promising role of handwriting analysis in various fields that necessitate age-related information.