/Photo-OCR

An application that detects hand-written characters using machine learning. Developed from scratch.

Primary LanguageJupyter Notebook

Photo-OCR

An application that detects hand-written numbers using machine learning. Developed from scratch.

Historically, the MNIST dataset is used as benchmark training and testing for several machine learning algorithms. This repository uses a simple neural network to classify the digits in these images with 97.5% accuracy.

MNIST Image

Problem-Statement

Given an image of a hand-written digit, detect the number written in the image.