A novel active contour model based on median absolute deviation for remote sensing river image segmentation. (August 2017)

• Record Type: Journal Article
• Authors: Han, Bin; Wu, Yiquan; Song, Yu
• Is Part Of: Computers & electrical engineering. Volume 62(2017)
• Page Start: 209
• Page End: 223
• Publication Date: 2017-08
• Subjects: Remote sensing river image -- Image segmentation -- Median absolute deviation -- Active contour model -- Region energy weights; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2017.04.005
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: The medians of pixel grayscale values inside and outside the curve are chosen as the region fitting centers instead of means, which can describe the intensity characteristics better. The external energy constraint terms of our model are defined by the median absolute deviation instead of the within-cluster variance, which characterizes the intensity differences of the object and background regions more accurately. The constant region energy weight is replaced by the fusion information of within-cluster variances and median absolute deviations of pixel grayscale values inside the object and background regions, which can be adaptively adjusted. The existing active contour models can not obtain accurate segmentation of the remote sensing river images. While our model outperforms the existing active contour models and can segment the remote sensing river images much more accurately and efficiently. Abstract: Aiming at the problem of the inaccurate segmentation of remote sensing river images by existing active contour models (ACMs), a novel ACM based on median absolute deviation for remote sensing river image segmentation is presented. Firstly, the external energy constraint terms of the presented model are defined by the median absolute deviation instead of the within-cluster variance in the Chan–Vese (CV) model. Secondly, in order to accelerate the evolution of the model, the fusion information of within-cluster variances and median absolute deviations of pixel grayscale values inside the object and background regions is utilized as the region energy weights. The corresponding experiments are carried out on a large number of remote sensing river images and the results illustrate that the presented model outperforms the existing ACMs, which can segment the remote sensing river images much more accurately and efficiently. Graphical abstract: