Spatial Fuzzy C-means Clustering Research Articles

An Integrated Algorithm of Spatial Fuzzy C-Means Clustering and Level Set for Indoor Scene Image Segmentation

Traditional fuzzy clustering algorithm is applicable for noiseless image segmentation. However, it is powerless for the images with noise, special point values and defects. An algorithm which combines spatial fuzzy clustering and level set for indoor scene segmentation is proposed in this paper. Firstly, the image is classified using fuzzy clustering with space information to get a larger difference in image gray level; secondly, the image is segmented using level set; finally, the contour in the boundary of target area is gotten accurately. The improved method can not only preserve details of images but also reduce the number of iterations. The results show that the proposed method has good segmentation quality and efficiency in segmentation for indoor scene image.

Multiphase Level Set-based Multi-objective Image Segmentation Cooperating with Spatial Fuzzy C-means

Multi-objective image segmentation is an important problem in machine vision. Multiphase level set model is sensitive to the initial contours of zero level set functions and image noise. Inappropriate initial position of zero level set or large image noise will lead to wrong segmentation results. A multiphase level set model-based multi-objective image segmentation method cooperating with spatial fuzzy C-means clustering,i.e. SFCM-MLS,is proposed. Coarse segmentation results are acquired with spatial fuzzy C-means clustering. Initial level set functions of multiphase level set model are defined with these coarse results in order to finish the accurate segmentation. The proposed SFCM-MLS cooperating image segmentation algorithm is tested on magnetic resonance image of human brain and CT image of human liver with tumours. Experiment results indicate that compared with classical MLS,SFCM-MLS algorithm is insensitive to initial contours of zero level set and improves the accuracy of multi-objective image segmentation.

Liver Segmentation from CT Image Using Fuzzy Clustering and Level Set

This paper presents a fully automatic segmentation method of liver CT scans using fuzzy c-mean clustering and level set. First, the contrast of original image is enhanced to make boundaries clearer; second, a spatial fuzzy c-mean clustering combining with anatomical prior knowledge is employed to extract liver region automatically; thirdly, a distance regularized level set is used for refinement; finally, morphological operations are used as post-processing. The experiment result shows that the method can achieve high accuracy (0.9986) and specificity (0.9989). Comparing with standard level set method, our method is more effective in dealing with over-segmentation problem.

An improved Spatial FCM algorithm Based on Artificial Bees Colony

In this paper, we presented a new spatial fuzzy clustering algorithm optimised by the Artificial Bees Colony (ABC) algorithm. By introducing the global search of ABC algorithm into the spatial Fuzzy C-Means Clustering algorithm, a new kind of methods (SFCM-ABC) is proposed. SFCM-ABC has two major characteristics. First it tackles better noisy image segmentation by making use of the spatial local information into the membership function. Secondly, it improves the global performance by taking advantages of the globalized search in the entire solution space of ABC rather than a localized searching like in FCM algorithm. Experiments with synthetic and real images show that SFCM-ABC is robust to noise compared to other methods. DOI: http://dx.doi.org/10.11591/ij-ai.v1i3.765 Full Text: PDF

Lip Processing and Modeling based on Spatial Fuzzy Clustering in Color Images

In this paper, a robust pre-processing algorithm based on spatial fuzzy clustering for a model based lip contour extraction is presented. This transformation aims at increasing discrimination between lips and skin to enhance the lip contour detection. For each pixel in the transformed data, a and b components of CIELAB color space are used as features in spatial fuzzy clustering phase. The employed clustering method takes into account both the distributions of data in feature space and the spatial interactions between neighboring pixels during clustering. Then, an elliptical and a Gaussian mask delete the pixels around lip that were previously clustered as lip pixels. Finally, a model is fitted upon the lip for extracting the lip contour. To show the effectiveness of the proposed method, pseudo hue segmentation and spatial fuzzy c-mean clustering were implemented and compared to our method. The results on VidTIMIT and M2VTS databases show that the novel scheme is able to extract the lip contour even in the situations that the nose and neck appeared in the lip region. Empirical results showed our method outperformed the present state-of-art methods.