Skip to content

Repository containing implemetation and documentation of bachelor's thesis Comparison of artificial intelligence methods for handwritten digit recognition at at Faculty of Electrical Engineering and Information Technology of Slovak University of Technology in Bratislava (FEI STU).

Notifications You must be signed in to change notification settings

FilipMikus/Comparison_of_artificial_intelligence_methods_for_handwritten_number_recognition

Repository files navigation

Bachelor’s thesis - Comparison of artificial intelligence methods for handwritten number recognition

Slovak University of Technology in Bratislava - Faculty of Electrical Engineering and Information Technology (FEI STU)

Abstract

Study Programme: Applied Informatics

Author: Bc. Filip Mikuš

Bachelor’s thesis: Comparison of artificial intelligence methods for handwritten number recognition

Supervisor: Ing. Eugen Antal, PhD.

Place and year of submission: Bratislava 2022

Digitization of historical documents is a phenomenon that allows the conservation of historical documents on digital media. Significant part of preserved documents also form encrypted documents with a specific structure - nomenclators. Encrypted historical documents are often transcripted to digital representation. This transcripted documents are then often part of complex cryptanalysis, too.

The aim of this work was to become familiar with structure and form of nomenclators, study, analyse and compare image preprocessing and handwritten text recognition techniques, and collect, analyse and compare suitable datasets. In the second part of our work, we implement a simple image preprocessing application and compare different implementations and configurations of handwritten number recognition systems.

Keywords: nomenclator, digitization of historical documents, image preprocessing, handwritten number recognition, machine learning, computer vision


Bakalárska práca - Porovnanie metód umelej inteligencie na rozpoznávanie rukou písaných číslic

Slovenská technická univerzita v Bratislave - Fakulta elektrotechniky a informatiky (FEI STU)

Súhrn

Študíjny program: Aplikovaná informatika

Autor: Bc. Filip Mikuš

Bakalárska práca: Porovnanie metód umelej inteligencie na rozpoznávanie rukou písaných číslic

Vedúci záverečnej práce: Ing. Eugen Antal, PhD.

Miesto a rok predloženia práce: Bratislava 2022

Digitalizácia historických dokumentov je fenomén, ktorý umožňuje uchovávanie historických prameňov na digitálnych médiach. Nezanedbateľnú časť takýchto dokumentov tvoria aj šifrované dokumenty špecifickej štruktúry, nomenklátory. Šifrované dokumenty sú často podrobené komplexným kryptologickým analýzam, s čím je spojená snaha o ich automatizovanú transkripciu.

Cieľom našej práce bolo oboznámiť sa so štruktúrou a podobou nomenklátorov, naštudovať, analyzovať a porovnať techniky na predspracovanie obrazu a rozpoznávanie rukou písaného textu, a zhromaždiť, porovnať a analyzovať vyhovujúce dátové množiny. V praktickej časti sme mali za úlohu podľa získaných poznatkov implementovať jednoduchú aplikáciu na predspracovanie obrazu a vykonať porovnanie rôznych implementácii a konfigurácii modelov na rozpoznávanie rukou písaných číslic.

Kľúčové slová: nomenklátor, digitalizácia historických dokumentov, predspracovanie obrazu, rozpoznávanie rukou písaných číslic, strojové učenie, počítačové videnie

About

Repository containing implemetation and documentation of bachelor's thesis Comparison of artificial intelligence methods for handwritten digit recognition at at Faculty of Electrical Engineering and Information Technology of Slovak University of Technology in Bratislava (FEI STU).

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published