REAL TIME DATA IMPORT FROM WEB POKER CLIENT

Boštjan Krajnc (2014) REAL TIME DATA IMPORT FROM WEB POKER CLIENT. EngD thesis.

Preview

Abstract

The diploma thesis presents development of application, which gathers data from currently opened on-line poker games in real time. It presents procedures for locating picture data and transforming this data into data structures, which can later be used by poker robot to make a decision. Procedure includes many steps. It starts with finding opened poker tables and capturing corresponding pictures of this tables. In next step edge detection and contour finding operations are applied to find some of the data. Pictures of data to convert needs to be appropriately preprocessed using techniques like binarization, resampling, with use of filters and morphological operations. Good preprocessing is the most important step because it has the biggest impact on optical character recognition accuracy. We used Tesseract OCR engine to do character recognition. Methods to detect and repair errors produced by OCR are also presented. Program was tested with 149 pictures of on-line poker games. Results are relatively good, overall accuracy of data recognition is 99,5%. Ideas to improve program are also given. Regarding usability, speed and even better quality of text recognition.

Item Type:

Thesis (EngD thesis)

Keywords:

OCR, on-line poker, computer vision, edge detection, contour finding, picture preprocessing, Open CV, Tesseract

Number of Pages:

Language of Content:

Slovenian

Mentor / Comentors:

Name and Surname	ID	Function
doc. dr. Peter Peer	294	Mentor

Link to COBISS:

http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=10505044)

Institution:

University of Ljubljana

Department:

Faculty of Computer and Information Science

Item ID:

2452

Date Deposited:

25 Mar 2014 09:23

Last Modified:

02 Apr 2014 13:07

URI:

http://eprints.fri.uni-lj.si/id/eprint/2452

Actions (login required)

View Item