Document Image Converter to XML File, Case of Tifinagh

Abstract

The document image converter is a necessity in our life for many reasons such as digitization of data in paper to secure them and gain time by automating this task, on the other hand preserving the environment by maintaining trees the first source of paper, for that this works based in a bi-cubic method for physical structure analyze and a graphs models representation for the characters recognition to generate at the end a standard XML file that can be used to create file is realised ( doc, pdf ,html…).

Characters recognition using keys points and convolutional neural network

“…In this current paper, we first give a review of others proposed tifinagh characters recognition systems developed by others searchers such as, Bencharef et al [1] used in feature extraction the riemannian metric these metric descriptors are reliable toward change scale. Boutaounte et al [2] are using the simplest possible geometric elements (segments and key points) to create the graph and extracting the impact matrix [3] for recognition. Essaady et al [4] who give a proposition used an analytical method used freeman coding for feature extraction and finite automation in recognition step.…”