Fork of tess-two rewritten from scratch to support latest version of Tesseract OCR.

A naive ocr that uses Artificial Neural Network built from scratch.

Automate the KYC Process using OCR (Implemented from scratch)

OCR using a simple network developed from scratch on NIST36 dataset vs with CNN on PyTorch on EMNIST dataset

A shell script to download and build tesseract-ocr from scratch

Generalised neural network implemented from scratch in python to teach step-wise functioning of a neural network and back-propagation training algorithm for optical character recognition (OCR)

🧠 OCR Neural Network A handwritten digit recognition system built from scratch using a simple feedforward neural network. Accepts 28×28 pixel grayscale inputs and classifies them into digits (0–9). Includes training, prediction, and a web-based drawing canvas.

This project implement basic OCR for Vietnamese from scratch with Pytorch, using CNN and BidirectionalLSTM

Implementation of OCR from scratch using Python and TensorFlow

No description available

easy OpenCV OCR implementation from scratch

No description available

OCR layout recognition model made from scratch using RESNET-UNET, UNET, CNN-RNN Hybrid, and Transformer for historical test recognition

Optical Character Recognition website

Classification of handwritten digit images into one of 10 classes representing integer values using tensorflow

No description available

OCR module from Scratch (Attention based OCR)

C++ neural network for OCR from scratch

Fine tuning keras-ocr python package with custom synthetic dataset from scratch

An OCR neural network coded from scratch using C, using stb_image.h for image handling

Graduation Project, a fully fledged DMS, supporting multiple users, in-document full text search, and a built-from-scratch OCR

This is the code written from scratch to implement a Neural network which can identify Hand written text using OCR data.

Dataless OCR: A neural network to recognize digits, written from scratch in C# .NET, and trained without any (external or labeled) data.

This repository contains code for a project that uses MediaPipe for hand landmark detection and a custom OCR system trained from scratch. The project aims to recognize handwritten text in the air using hand gestures captured through a camera.

A unified platform where users can upload documents (PDF, images, contracts, insurance policies, academic notes, receipts), and my backend exposes an API that allows the frontend to query, summarize, extract insights, verify information, and compare documents—powered by ML, OCR, embeddings, built from scratch

A collection of neural network tutorials and implementations across multiple languages. Includes building perceptrons, multilayer perceptrons, CNNs, and LSTMs from scratch in Python, Go, F#, C#, and JavaScript. Applications like OCR, traffic sign recognition, and music generation, neural network fundamentals and framework internals (SlowTorch).

This project demonstrates a simple yet effective Optical Character Recognition (OCR) system using a neural network implemented from scratch. The project includes components for training a neural network on the MNIST dataset and a web-based interface that allows users to draw digits and get real-time predictions.

No description available

No description available

I tried to code up an character recognition model only by using base python.