Search

A local search-based version of the so-called genetic sequential image segmentation (GeneSIS) algorithm is presented in this paper, for the classification of remotely sensed images. The new method combines the properties of the GeneSIS framework with the principles of the region growing segmentation algorithms. Localized GeneSIS operates on a fine-segmented image obtained after preliminary watershed transformation. Segmentation proceeds by iterative expansions emanating from object cores, i.e., connected components of marked watersheds. At each expansion trial, the process involves three successively performed operations: 1) generation of the object’s neighborhood to a specified order; 2) local exploration of the neighborhood through an evolutionary algorithm to identify the best expansion to be merged; and 3) rearrangement of the object neighborhoods. We propose two priority strategies for the selection of the objects to be expanded and two different modes of operation performing either supervised or semisupervised segmentation of the image. The combination of the priority strategies and segmentation modes lead to four different implementations of localized GeneSIS. Due to the local search approach adopted here, the resulting algorithms have considerably lower execution times, while at the same time, they provide comparable classification accuracies compared to those produced by previous GeneSIS variants. Experimental analysis is conducted using a hyperspectral forest image, a multispectral agricultural image, and the Pavia Centre image over an urban area. Comparative results are also provided with existing segmentation algorithms. [ABSTRACT FROM PUBLISHER]

In this paper, we propose an integrated framework of the recently proposed Genetic Sequential Image Segmentation (GeneSIS) algorithm. GeneSIS segments the image in an iterative manner, whereby at each iteration, a single object is extracted via a genetic algorithm-based object extraction method. This module evaluates the fuzzy content of candidate regions, and through an effective fitness function design provides objects with optimal balance between fuzzy coverage, consistency and smoothness. GeneSIS exhibits a number of interesting properties, such as reduced over-/undersegmentation, adaptive search scale, and region-based search. To enhance the capabilities of GeneSIS, we incorporate here several improvements of our initial proposal. On one hand, two modifications are introduced pertaining to the object extraction algorithm. Specifically, we consider a more flexible representation of the structural elements used for the object's extraction. Furthermore, in view of its importance, the consistency criterion is redefined, thus providing a better handling of the ambiguous areas of the image. On the other hand we incorporate three tools properly devised, according to the fuzzy principles characterizing GeneSIS. First, we develop a marker selection strategy that creates reliable markers, particularly when dealing with ambiguous components of the image. Furthermore, using GeneSIS as the essential part, we consider a generalized experimental setup embracing two different classification schemes for remote sensing images: the spectral-spatial classification and the supervised segmentation methods. Finally, exploiting the inherent property of GeneSIS to produce multiple segmentations, we propose a segmentation fusion scheme. The effectiveness of the proposed methodology is validated after thorough experimentation on four data sets. [ABSTRACT FROM PUBLISHER]