MNIST-Fraction
收藏资源简介:
MNIST-Fraction数据集是由肯特州立大学团队开发,旨在用于手写数学分数的识别与分析。该数据集基于著名的MNIST数据集,通过合成方法生成,包含72159条手写分数图像,涵盖单双位数的分子和分母。数据集的创建过程包括从MNIST数据集中提取数字图像,结合自定义的分数条设计,生成多样化的分数表示。MNIST-Fraction数据集主要应用于数学教育领域,旨在通过AI技术提升分数识别的准确性,从而支持自动化评分系统和教育技术平台,帮助学生更好地理解和掌握分数这一基础数学概念。
The MNIST-Fraction dataset was developed by a team from Kent State University, aiming to be used for the recognition and analysis of handwritten mathematical fractions. Built upon the renowned MNIST dataset, this dataset is generated through synthetic methods, containing 72159 handwritten fraction images with single-digit and double-digit numerators and denominators. The dataset creation process includes extracting digit images from the MNIST dataset and combining them with custom-designed fraction bars to generate diverse fraction representations. The MNIST-Fraction dataset is mainly applied in the field of mathematics education, with the objective of enhancing the accuracy of fraction recognition via AI technologies, thus supporting automated grading systems and educational technology platforms, and helping students better understand and master the fundamental mathematical concept of fractions.

- 1MNIST-Fraction: Enhancing Math Education with AI-Driven Fraction Detection and Analysis肯特州立大学 · 2024年



