Optical Characters Recognition

EPITA S3 Project (2023#) Optical character recognition software written in C.

Group Name: Trantor

Members:

• hugo.moreau (Moreau Hugo) => github
• anthony1.truong (TRUONG Anthony)
• maxime.chardon (CHARDON Maxime) => github

Neural Network in /NeuralNetwork

• contain file for build a neural network.
• contain a xor file too.
• We can save and load training from neural network.
• Usage:

$ make xor
$ ./xor #training
$ ./xor n #for n training

$ make ocr
$ ./ocr #train by default 1000 iterations
$ ./ocr <path_to_char> #compute the neural network on this char image

Segmentation in /segmentation

• contain file for cut an image in bloc, line and character.
• coutain file for applying grayscale and binarization.
• Usage

$ make grayscale
$ ./grayscale <path_to_image>

$ make segmentation
$ ./segmentation <path_to_image>

OCR

The program

$ make
$ ./main

• step1: choose an image
• step2: load the image (optionnal)
• step3: train the ocr for better performance
• step4: run the OCR