bensonruan/Hand-Written-Digit-Recognition

Leverages TensorFlow.js to run a pre-trained CNN model (trained on MNIST with Keras) directly in the browser, enabling real-time digit classification without server calls. Features an interactive canvas interface supporting both desktop and touch input, with prediction confidence visualized via Chart.js. The model uses convolutional and pooling layers followed by dense layers with dropout regularization for robust classification across handwritten digit variations.

No commits in the last 6 months.