سال انتشار: ۱۳۹۰

محل انتشار: پنجمین کنفرانس بین المللی پیشرفتهای علوم و تکنولوژی

تعداد صفحات: ۸

نویسنده(ها):

Mojtaba Sabet – Payame Noor University faculty of engineering, Tehran, Iran
Saeed Ayat – Payame Noor University faculty of engineering, Najafabad, Iran
Reza Askari Moghadam – Payame Noor University faculty of engineering, Tehran, Iran

چکیده:

The growing availability of inexpensive computer memory means that large amount of on-line images are now stored routinely, raising problems in providing fast and flexible access to images. Recently researchers proposed different methods of storage and retrieval of symbolic images but these methods have various shortcomings, for example most of them have high time complexity and are not suitable for dynamic image databases.In this paper a faster and more efficient method for storage and retrieval of symbolic images in databases is proposed. In this method TSR1 is used to represent spatial relationships existing among the elements of an image and a kind of binary neural network called CMM2 is used as a data structure. This proposed method has low time complexity and has the capability of different types of retrieval such as exact mach retrieval and partial match retrieval, it also has the ability to retrieve the transformed3 images and can be used in dynamic databases. The study made in this work reveals that this method bears various advantages when compared to other existing methods