| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

Show document Help

Title:OPTIČNO RAZPOZNAVANJE ZNAKOV Z NEVRONSKIMI MREŽAMI NA GRAFIČNI PROCESNI ENOTI
Authors:ID Furlan, Miha (Author)
ID Potočnik, Božidar (Mentor) More about this mentor... New window
ID Strnad, Damjan (Comentor)
Files:.pdf UNI_Furlan_Miha_2011.pdf (6,59 MB)
MD5: 8DE1CC37EE24C052E94FD7FAACD68137
PID: 20.500.12556/dkum/53a351eb-eee9-4015-b08d-c44bee884519
 
Language:Slovenian
Work type:Bachelor thesis/paper
Organization:FERI - Faculty of Electrical Engineering and Computer Science
Abstract:Cilj diplomske naloge je izdelava sistema za optično prepoznavanje znakov z uporabo nevronske mreže. Računsko zahtevne dele sistema smo pohitrili z uporabo grafične procesne enote (GPU). Naš sistem za OCR opišemo v treh glavnih sklopih: razbitje dokumenta na znake (segmentacija), prepoznavanje posamičnih znakov ter paralelizacija izvajanja na GPU. Zatem predstavimo aplikacijo, v katero smo integrirali našo rešitev. Rezultati testiranj so pokazali, da je natančnost prepoznavanja znakov OCR-A in OCR-B okrog 98%, Courier New pa 92%, medtem ko je pohitritev izvajanja kode na GPU bila minimalno petkratna napram izvajanju na CPU.
Keywords:optično prepoznavanje znakov, nevronske mreže, grafična procesna enota
Place of publishing:Maribor
Publisher:[M. Furlan]
Year of publishing:2011
PID:20.500.12556/DKUM-19835 New window
UDC:004.352.242:004.032.26(043.2)
COBISS.SI-ID:15317014 New window
NUK URN:URN:SI:UM:DK:9PYJHR2T
Publication date in DKUM:12.09.2011
Views:2445
Downloads:187
Metadata:XML DC-XML DC-RDF
Categories:KTFMB - FERI
:
Copy citation
  
Average score:(0 votes)
Your score:Voting is allowed only for logged in users.
Share:Bookmark and Share



Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Secondary language

Language:English
Title:OPTICAL CHARACTER RECOGNITION BY USING NEURAL NETWORKS ON GRAPHICAL PROCESSING UNIT
Abstract:The goal of this diploma work is to develop a system for optical character recognition (OCR) by using neural network. Computationally intensive parts of the system are going to be implemented on the graphics processing unit (GPU). We present our OCR system in three main parts: segmentation of document on characters, recognition of individual characters, and parallelization of execution on the GPU. Afterwards, we present an application with integrated our solution. Results of testing pointed out that the accuracy of OCR-A and OCR-B characters recognition was around 98%, while at Courier New characters this rate was 92%. A code execution on GPU was at least five times faster than on CPU.
Keywords:Optical character recognition, neural networks, graphical processing unit


Comments

Leave comment

You must log in to leave a comment.

Comments (0)
0 - 0 / 0
 
There are no comments!

Back
Logos of partners University of Maribor University of Ljubljana University of Primorska University of Nova Gorica