Recognizing Preschool Children Handwriting Using CNN

Abstract

The major difficulty of identifying handwritten numbers and alphabetic letters produced by preschoolers is addressed in this study. It can be challenging for conventional automated recognition systems to correctly interpret children’s handwriting at this early developmental stage because it is frequently irregular unstructured, and inconsistent in shape, size, and orientation. Convolutional neural networks (CNNs), a kind of deep learning model well-known for being effective in image classification tasks, are used in this study to get around these difficulties. The CNN’s design focuses on managing the variability common in preschool writing while classifying both numbers (0–9) and both uppercase and lowercase letters (A–Z, a-z). The aim is to create a model that can reliably identify characters even when the handwriting is inconsistent. Through more effective literacy development monitoring, this system seeks to offer early childhood educators useful assistance. Automating the recognition process allows teachers to spot students who might require more help early on, allowing for individualized instruction and prompt intervention. The experimental outcome shows that for recognizing hand written numbers as well as uppercase and lowercase letters written by preschool children, the designed CNN is capable of achieving a 86% accuracy. The goal of this research is to improve early learning assessment by bridging artificial intelligence and education.