期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2011
卷号：2
期号：3
页码：1267-1271
出版社：TechScience Publications
摘要：Object detection for computer vision is one of the key factors for image understanding. It is still a challenge today to accurately determine an object from an image where similar shaped objects are present in a large number. An object can be defined as a structure within a document that has a particular meaning in the context of the application. An object can be a symbol, character or any disconnected component. The text which usually forms major information provider for the image is one of the crucial events to be made successful. Although there is no generalized algorithm for this problem, this paper helps in understanding the various steps that are taken to recognize the text in an image and the various methods that are implemented in each step to go ahead in the process. Contour based extraction is the crucial step in this process. It also proposes the best method that gives the most appropriate results for a given input setw8{
