Mathematically Modified Deep Learning Model Assisted Handwritten Digit Recognition for Intelligent Document Processing Systems
Ali A Ibrahim Alasadi, S H Manjula, Aseel Smerat, Komali Govindu, G. Sivakumar, V. Sahasranamam
Abstract
Information processing requires handwritten digit recognition; however, methods of writing and image defects, such as brightness changes, blurring, and noise, make image recognition challenging. This paper presents a strategy for categorizing offline handwritten digits in Devanagari script and Roman script (English numbers) using Deep Learning (DL), a branch of Machine Learning (ML) that utilizes Neural Networks (NN) with multiple layers to attain classified representations of input autonomously. The research study develops classification algorithms for recognizing handwritten digits in the numerical characters (0–9), analyzes combination approaches for classifiers, and evaluates their accuracy. The study aims to optimize recognition results when working with multiple scripts simultaneously. It proposes a simple profiling method, Linear Discriminant Analysis (LDA) implementation, and an NN structure for numerical character classification. However, testing shows inconsistent outcomes from the LDA classifier. The approach, which combines profile-based Feature Extraction (FE) with advanced classification algorithms, can significantly improve the field of HWR numerical characters, as evidenced by the diverse outcomes it produces. The model achieved an accuracy of 98.98% on the MNIST dataset. In the CPAR database, this work conducted a cross-dataset evaluation with an accuracy of 98.19%.