Swarm optimization based texture classification in extreme scale variations

Author:

Kumar K. Santhosh

Abstract

Texture analysis has remained a remarkably trendsetting and productive field of research in the last two decades. There has been much progress, but the impact of illumination changes on automated texture classification and segmentation has gained very less focus. Research work carried out in the field of texture identification frequently focuses on the identification of textures with intraclass changes including illumination, rotation, viewpoint and small scale variations. Consequently, variations in texture owing to modifications in scale constitute among the ones that are difficult to manage. In this research work, as the first step, the classification of textures due to vast changes in scale is studied. In order to deal with this problem, first the solution is introduced and then the scale changes are reduced based on the predominant patterns in the texture. Inspired by the challenges imposed by this issue, a novel swarm intelligence approach known as Ant Colony Optimization (ACO) algorithm is introduced for modifying the components in the hidden layers used during the network training, for the extraction of more useful semantic texture patterns.

Publisher

Universidad Tecnica de Manabi

Subject

Education,General Nursing

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3