SOM Surface Feature Extraction Techniques for Retrieval of Multimedia Data: Designed and Implemented Model

Abstract

This paper represents an retrieval system for images (IRS) based on SOM technique using it as a method for fetching images identical with a query image from an image database. This mechanism is based on understanding the user’s perspective in selecting form the different images are similar to each other. The image queries are performed through MATLAB GUI and the images are iteratively retrieved in the form of output as long as there are matching images in the database. Proposed the system based on SOM techniques which are implemented in MATLAB. This will be the useful for making a system to take the selected images from the large database. Created, loaded and deleted after the image selection. Matching of the image from the database with the input image through applying the various distance function from the neighboring database and rearranged again so that all the things are easily mapped by the choosing closest images every time. For analyzing the system, prepared a model for the best result/retrieving the images in very less time. This will be helpful for everyone those are working on the selection of images from the large dataset.