Dataset Creation Framework for Handwritten Exponent Recognition

We introduce a novel framework for creating datasets specifically tailored to advance the recognition of handwritten mathematical expressions, with a focus on handwritten exponents. Building upon the MNIST dataset, our framework facilitates the generation of images featuring base digits (0-9) combined with handwritten exponents. The process involves systematic resizing, positioning, and overlaying of digits and exponents on a standardized black canvas to ensure uniformity. Each image is accompanied by detailed metadata, which is systematically recorded in a CSV file. This framework provides a structured approach to dataset creation, offering a valuable resource for developing and testing machine learning models designed to recognize complex handwritten mathematical expressions. The dataset created using this framework holds promise for enhancing applications in educational technology and digital documentation, where accurate interpretation of mathematical handwriting is crucial.