TensorFlow.js MNIST Image Classifier

This project is a web-based digit classifier built using TensorFlow.js and the MNIST dataset. It classifies hand-drawn digits (0-9) by predicting the number displayed in the image input. The application leverages TensorFlow.js to load, process, and classify images in real time within the browser.

Table of Contents

• Overview
• Features
• Demo
• Installation
• Usage
• Project Structure
• Technical Details
• Credits
• License

Overview

This project utilizes the TensorFlow.js library to load a pre-trained model that can classify images of handwritten digits from the MNIST dataset. The classifier predicts the digit displayed in a 28x28 pixel grayscale image, with correct predictions highlighted in green and incorrect predictions in red.

Features

• Real-time prediction: Provides instant feedback on digit predictions.
• Color-coded predictions: Green text indicates correct predictions, while red text indicates incorrect predictions.
• Browser-based model: Built entirely in JavaScript using TensorFlow.js, making it platform-independent and easy to use.

Demo

You can view the live demo of the project at Your Demo Link Here.

Installation

Prerequisites

• Ensure you have a modern web browser that supports JavaScript and TensorFlow.js.
• Internet connection to load TensorFlow.js library and model data.

Steps

1. Clone this repository:

   git clone https://github.com/bozics_chucky/tensorflowjs-mnist-classifier.git
   cd tensorflowjs-mnist-classifier