文章基本信息

标题：Interactive Image Segmentation Using Combined MRF and Ant Colony Optimization
本地全文：下载
作者：Vrushali D. Mendhule ; Gaurav Soni ; Alesh Sharma 等
期刊名称：International Journal of Engineering and Computer Science
印刷版ISSN：2319-7242
出版年度：2015
卷号：4
期号：6
页码：12281-12288
出版社：IJECS
摘要：Image segmentation is the process that partitions an image into region. Although many literatures studied automated imagesegmentation, it is still difficult to segment region-of-interest in any kind of images. Thus, manual delineation is important yet. [2] Inorder to shorten the processing time and to decrease the effort of users, this paper introduces the approaches of interactive imagesegmentation method based on MRF and Ant colony optimization. In this paper we proposed a segmentation algorithm combined MRFwith ACS, which not only applied ACS as optimization algorithm but also introduced the neighborhood pheromone interaction rules intoACS under MRF model. Interactive segmentation aims to separate an object of interest from the rest of an image. This problem incomputer vision is known to be hard, and very few fully automatic vision systems exist which have been shown to be accurate and robustunder all sorts of challenging inputs. Most of the previous works require users to trace the whole boundary of the object. When theobject has a complicated boundary, or the object is in a highly textured region, users have to put great effort into iteratively correctingthe selection. [1] Dirichlet Process Multiple-View Learning (DPMVL) for image segmentation technique produces very effectivesegmentation results as compare to previously existing techniques. DPMVL use MRF model for smoothing the segmentation. This can befurther improved by using MRF-based image segmentation using Ant Colony System which works effectively and provide an alternativecomputational algorithm for building interactive image editing tools. In this paper, we present an interactive segmentation frameworkthat integrates image appearance and boundary constraints in a principled way using combined MRF and ant colony optimization. Wehave improved proposed technique by using modified technology which have more interactivity, user control of segmentation process,and reach a satisfied result among the noise restraint, edge preservation and computation complexity. Experimental results are providedto demonstrate the superior performance of the proposed approach. A comparison with other standard operators is also discussed andthe proposed method produced acceptable results within reasonable amounts of time. It is shown that the proposed algorithm based onant colony optimization and MRF achieves better performance compared to the typical interactive image segmentation methods withoutusing ant colony optimization concept.
关键词：Image Processing; Image segmentation; Interactive image segmentation; DPMVL; MRF; Ant colony optimization