This paper proposes a methodology to obtain a fully automatic color segmentation algorithm based on the Normalized Cut (Ncut) proposed by Shi and Malik, using recent findings in color perception. A weighting matrix computed using a perceptually uniform color space (CIE $$L^*a^*b^*$$ L ? a ? b ? ) and color distance formulae correlated with the visually perceived color differences (CIE94 and CIEDE2000); a stopping condition related to perceptual criteria; an automatic parameters setting required to compute the affinity matrix are proposed. To test the proposed methodology, a wide study about the influence of the color space choice, different stopping conditions, and different similarity measurements is carried out. These alternatives are exhaustively evaluated using perception-related measurements (S-CIELAB) and general segmentation evaluation metrics applied to the 500 images of the Berkeley database. The results showed that the proposed method outperforms Ncut based on other color spaces, similarity measure or stopping conditions. Furthermore, the usability of the method is increased by replacing the manual parameter setting for an automatic.

Normalized Cut optimization based on color perception findings. A comparative study

Image segmentation is an important process in Melanoma diagnosis. It divides the image into segments which provides a better path for extracting features and classifying them accordingly and so far many literatures have explained how the segmentation has been carried out for gray scale images. Though most algorithms have been written for gray scale images it is not necessary that we must stick on only to that instead we can also use color images directly for segmentation process. On doing so, variation in color can be used as an important feature for melanoma diagnosis. Color is a unique feature which can be used to provide a distinct differentiation between Melanoma and Benign. Hence on applying this color image segmentation, earlier detection can be done easily. This literature presents a survey on existing techniques for color image segmentation. Color images when segmented directly, yields better differentiation between the lesions.

A Survey on Color Image Segmentation Techniques for Melanoma Diagnosis

Image segmentation is an important research topic in the field of computer vision. Spectral Clustering (SC) algorithm is one of the most popular used clustering methods for image segmentation. However, the cluster number must be estimated by expertise users to be determined. This limits its application in image segmentation. In this paper, we proposed an image segmentation method based on saliency map and spectral clustering algorithm (SC) that includes the initialization and determination of the number of cluster k. Quantitative and qualitative experimental evaluation on MSRA-1000 public image dataset depicts the promising results from the proposed method by outperforming four state of the art image segmentation methods.

An object segmentation method based on saliency map and spectral clustering

Image segmentation is one of the most crucial techniques in computer vision. In this paper, we propose an improved approach for color image segmentation, Texture clustering based JSEG (T-JSEG). The T-JSEG combines the texture information with the classical JSEG algorithm. We use the Gabor feature clustering to produce texture map and combine it with color map to achieve a new map for J value computing. The color image segmentation result is improved by adding the texture information to the JSEG. Experiments on natural color images of Berkeley Segmentation Database verify that our method is very effective on color image segmentation and more accurate boundaries of regions are found in comparison with the classical JSEG algorithm.

Image segmentation with texture clustering based JSEG

Image Segmentation has been considered as challenging task in image processing field. In this paper we proposed a new method for image segmentation purpose. Our proposed method is based on combining Grasshopper Optimization Algorithm (GOA) and k-means clustering. Grasshopper Optimization Algorithm (GOA), inspired by nature, divides the search process into two parts of exploration and utilization. The proposed algorithm in terms of mathematics models the behavior of Grasshopper groups in nature for solving the problems of optimization. We evaluate the proposed algorithm function based on eight-image segmentation and compare the obtained results of the tests of PSNR scale and SSIM measurement with four moth-flame whale algorithms by harmony search and social spider. The results indicate that the proposed algorithm with a simple threshold gives us a more optimized response regarding other state of arts in terms of both objective and subjective criteria.

Multilevel Image Segmentation using Hybrid Grasshopper Optimization and k-means Algorithm

Image segmentation is the first step in computer vision project. But it is also the key procedure from image processing to image analysis. With the continuous development of computer hardware and technology advances, more attention has been paid to the color image. Based on previous image segmentation technology, this paper proposes a novel color image segmentation method. The method improves the JSEG(Joint Systems Engineering Group) algorithm, and it uses the results of JSEG as the input of Ncut(Normalized Cuts). Experiments show that this proposed algorithm can obtain a good segmentation results. The two main contributions of this paper are: 1) it is effective to overcome the problems of the large computation, 2) the final image segmentation boundary is more accurate.

Li Wang

Papers

A novel color image segmentation algorithm based on JSEG and Normalized Cuts