An Efficient and Scalable Technique for Texture Analysis to Obtain in IR Model

Abstract

Nowadays the popularity of multimedia contents search is increasing rapidly such as images, videos, and the audio. Among them, the images are much popular for search and data retrieval processes. In literature there are a number of techniques are exist for efficient and precise image retrieval but according to the conclusion of literature the content based techniques are a much promising approach. In content-based image retrieval systems, the image internal descriptors or features are used for representing the image contents. In these descriptors the shape, colour and texture are primary image content descriptors. A number of works are available for different combinations of these features but there is work exist for the texture analysis. In order to recognize the feature based similar objects, the texture feature played an essential role. In this work the texture analysis is a key area of interest, therefore different texture analysis techniques are studied and LBP is selected for further work. Thus the LBP based texture classification technique is proposed and implemented in this work first. In addition of to demonstrate the effectiveness of the proposed texture analysis technique, the application of the approach in content-based image retrieval is also demonstrated. The implementation of the proposed approach is provided using the visual studio technology. After implementation of the technique, the performance of the system is evaluated in terms of precision, recall, and f-measures. Additionally, for finding the computational complexity the time and space complexity is also computed and compared with the traditional LBP based technique. According to the obtained performance, the proposed technique results from more precise results in less resource consumption.