Embedded Development of Handwritten Digit Recognition

Intro

• This is about the course project of 'Embedded System' at Tongji.
• We realize 📝handwritten digit recognition👀 based on STM32H747I-DISCO.
• More details can be found in Report.pdf

Demo

• The function is shown as below.

Model

• We use a basic CNN structure to detect and classify digits, which performs well on the MNIST Dataset.
• Look at Folder 'Handwritten Digit Recognition_Python' for more details.

Embedded System

• We use CubeAI as a bridge between neural networks and embedded systems.
• Look at Folder 'Project' for more details.