Alireza Akoushideh
University of Twente, Netherlands
Title: A New Approach to Improve the Throughput of the Texture Classification
Biography
Biography: Alireza Akoushideh
Abstract
In this speech, we talk about features' value range (FR) approach to enhance the throughput of texture classification. The proposed approach is used as a pre-classifier with a simple construction. The FR approach selects a few categories as class candidates for an input image and reduces the comparison time of the main classifier. The proposed approach works based on gathering information from local areas. We define two configuration parameters and some criteria for evaluation of the FR approach in different configurations. We discuss about the performance of the FR in presence of different levels of Gaussian noise. Finally, we propose two approaches to improve the efficiency of the FR approach