Abstract
This paper presents a novel variational model based on fuzzy region competition and statistical image variation modeling for image segmentation. In the energy functional of the proposed model, each region is characterized by the pixel-level color feature and region-level spatial/frequency information extracted from various image domains, which are modeled by the windowed bit-plane-dependence probability models. To efficiently minimize the energy functional, we apply an alternating minimization procedure with the use of Chambolle’s fast duality projection algorithm, where the closed-form solutions of the energy functional are obtained. Our method gives soft segmentation result via the fuzzy membership function, and moreover, the use of multi-domain statistical region characterization provides additional information that can enhance the segmentation accuracy. Experimental results indicate that the proposed method has a superior performance and outperforms the current state-of-the-art superpixel-based and deep-learning-based approaches.
Highlights
Image segmentation, which aims to partition an image into homogeneous regions [1,2,3], is one of the most challenging problems in computer vision and has various applications, such as pattern recognition and medical imaging
The following are some discussions about the algorithms and main points observed from the figures
This may be explained by the fact that fuzzy region competition with Gaussian mixture model (FRCGMM) performs highly precise fitting of data with Gaussian mixture model, and that L1 fuzzy segmentation (L1FS) introduces two sets of auxiliary variables and requires solving a collection of sub-problems based on the alternating direction methods of multipliers
Summary
Image segmentation, which aims to partition an image into homogeneous regions [1,2,3], is one of the most challenging problems in computer vision and has various applications, such as pattern recognition and medical imaging. Chan and Vese [9] studied a particular case of the Mumford-Shah model using the piecewise constant functions and applied their model to a two-phase image segmentation problem They proposed a curve evolution technique with a level set [10]. In [29], the authors proposed an improved fuzzy region competition-based framework via the hierarchical strategy so that the minimization problem is always convex during the iterative calculation Their method was applied to noisy SAR images, with satisfactory segmentation results. A characteristic GGD based on Kullback-Leibler divergence [35] and the generalized Gamma density [36] were proposed and applied to supervised texture classification with promising results While these models usually work well for most distributions of image variation, they always assume the distributions have a specific structure (such as symmetry, monotone and periodicity) and cannot model fluctuating distributions. Incorporating these models to characterize image regions into the energy functional would help to enhance the segmentation of texture images and deal with challenging segmentation problem
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
