Exploring CNN-Based Recognizers for Handwritten Script

Handwritten Character Recognition is a critical domain in the field of Artificial Intelligence (AI), playing a significant role in the automation of document digitization and character evaluation. This paper focuses on the use of Convolutional Neural Networks (CNNs) for handwritten digit and character recognition, with particular emphasis on accuracy and computational efficiency. The proposed CNN model achieves 99.70 % accuracy on the MNIST dataset and 92.91% on the NIST dataset. A comparative evaluation with traditional neural network architectures, such as Multilayer Perceptions (MLPs) and Recurrent Neural Networks (RNNs), highlights CNNs as a superior architecture for processing spatial information. Additionally, this research explores the implications of handwriting recognition for identifying learning disabilities and discusses potential applications in educational contexts and assistive technologies.