No abstract

No abstract

Snake, active contour or deformable active contour has been widely used in medical image segmentation area. In this paper, comparison between Gradient Vector Flow (GVF) snake and Enhanced Distance (ED) snake in segmenting microcalcifications is carried out. The performance is measured based on actual area of the average percentage difference traced by expert radiologists. Results obtained shows that the values of average percentage difference for the GVF and ED snake are 4.3% and 6.68% respectively. These results indicate that the GVF snake has better performance with 95.7%.

No abstract

No abstract

No abstract

1Selective image segmentation is the task of extracting one object of interest among 2 many others in an image based on minimal user input. Two-phase segmentation models 3 cannot guarantee to locate this object, while multiphase models are more likely to classify 4 this object with another features in the image. Several selective models were proposed 5 recently and they would find local minimizers (sensitive to initialization) because non-convex 6 minimization functionals are involved. Recently, Spencer-Chen (CMS 2015) has successfully 7 proposed a convex selective variational image segmentation model (named CDSS), allowing 8 a global minimizer to be found independently of initialisation. However, their algorithm is 9 sensitive to the regularization parameter µ and the area parameter θ due to nonlinearity in 10 the functional and additionally it is only effective for images of moderate size. In order to 11 process images of large size associated with high resolution, urgent need exists in developing 12 fast iterative solvers. In this paper, a stabilized variant of CDSS model through primal-dual 13 formulation is proposed and an optimization based multilevel algorithm for the new model 14 is introduced. Numerical results show that the new model is less sensitive to parameter µ 15 and θ compared to the original CDSS model and the multilevel algorithm produces quality 16 segmentation in optimal computational time. 17 AMS subject classifications: 62H35, 65N22, 65N55, 74G65, 74G75 18Image segmentation is a fundamental task in image processing aiming to obtain meaningful 22 partitions of an input image into a finite number of disjoint homogeneous regions. Segmentation 23 models can be classified into two categories, namely, edge based and region based models; other 24 models may mix these categories. Edge based models refer to the models that are able to 25 drive the contours towards image edges by influence of an edge detector function. The snake 26 algorithm proposed by Kass et al. [33] was the first edge based variational model for image 27 segmentation. Further improvement on the algorithm with geodesic active contours and the 28 level-set formulation led to effective models [14, 49]. Region-based segmentation techniques try 29 to separate all pixels of an object from its background pixels based on the intensity and hence 30 find image edges between regions satisfying different homogeneity criteria. Examples of region-31 based techniques are region growing [30, 9], watershed algorithm [30, 10], thresholding [30, 53], 32 and fuzzy clustering [50]. The most celebrated (region-based) variational model for the images 33 (with and without noise) is the Mumford-Shah [43] model, reconstructing the segmented image 34 as a piecewise smooth intensity function. Since the model cannot be implemented directly and 35 easily, the Mumford-Shah general model [43] was often approximated. The Chan-Vese (CV) [21] model is simplified and reduced from [43], without approximation. The simplification is to 37 replace the piecewise smooth...

Edge detection has been widely used especially in medical image processing field. In this paper we are comparing Sobel, Prewitt and Laplacian of Gaussian (LoG) edge detection techniques in segmenting the boundary of microcalcifications. The edge detection must satisfy the breast phantom scoring criteria before the segmentation phase is carried out. Then, all of the edge detection techniques are implemented in the Enhanced Distance Active Contour (EDAC) model for the segmentation process. Results obtained from Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) curve shows that the Prewitt edge detection has the highest value of AUC, followed by the Sobel and LoG which are 0.79, 0.72 and 0.71 respectively.