A New Method Of Image Segmentation Based On Multi-feature

Image segmentation is an important tool in image processing and it can serve as an efficient front end to sophisticated algorithms and thereby simplify subsequent processing. With the development of segmentation technologies, image segmentation in image coding, pattern recognition, computer vision, medical image analysis and other fields have a practical application. The researchers have proposed a lot of algorithms, but these algorithms have some shortcomings. So far, image segmentation is still a challenging task.Recently, the segmentation methods focus on combining Multi - features and multi - segmentation technologies. In this paper, fuzzy clustering algorithm and support vector machine are carefully studied and summarized. After comprehensively reviewing the basic principles and existed methods, the author chooses the approaches based on features to solve this problem. In this paper, the author achieves effective and innovative FCM and SVM, which can be completed as follows:1. Fuzzy clustering algorithm has been studied in detail, especially fuzzy C-means algorithm. The initial parameters are seriously studied, such as the initial clustering center and so on. Due to the poor robustness of FCM, how to utilize the spatial information has been a key point.2. In the conventional FCM image segmentation algorithm, cluster assignment is based solely on the distribution of pixel attributes in the feature space, and does not take into consideration the spatial distribution of pixels in an image. In this paper, we present a novel FCM image segmentation scheme by utilizing local contextual information and the high inter-pixel correlation inherent. Experimental results showed the proposed method achieves competitive segmentation results compared to other FCM-based methods, and is in general faster.3. The kernel function, kernel parameters, the penalty factor and other factors on the support vector machine (SVM) are considered. The proposed unsupervised SVM method provides a reference for image segmentation.4. We present a pixel-based color image segmentation using Support Vector Machine (SVM) and Fuzzy C-Means (FCM). This image segmentation not only can be fully taken advantage of the local information of color image but also the ability of SVM classifier. Experimental evidence shows that the proposed method has a very effective computational behavior and effectiveness, and decreases the time and increases the quality of color image segmentation in compare with the state-of-the-art segmentation methods recently proposed in the literature.