BACKGROUND: Coronary artery image segmentation is an important auxiliary method for coronary artery disease diagnosis. OBJECTIVE: The classical region growing algorithms, which only consider the intensity of pixels, are noise-sensitive and require manual interaction. To this end, recent methods simultaneously consider the intensity of pixels and multi-scale analysis with the region growing. Nevertheless, these methods are not fully optimized and they suffer from the drawbacks of over- or under-segmentation in many cases. METHODS: In this paper, we propose a region growing based coronary artery segmentation method. Different from the existing methods, the variable sector search area is considered in the region growing technique. A growing rule is proposed to segment the vessel, which combines the Hessian vector and the region growing with the variable sector search area. To further improve the quality of segmentation, we propose an optimization of removing some small disconnected regions. RESULTS: Our proposed method can search more branches while segmenting the vessel, even the small ones. It keeps an acceptable performance when dealing with stenosis and large curvature of blood vessels. CONCLUSIONS: Quantitative evaluations are conducted on coronary angiography and the results show that the proposed method achieves a higher DSC ratio and a more reliable sensitivity ratio.

Coronary artery segmentation based on the Computerized Tomography Angiography images is an important subject in the field of vascular medical imaging. In this paper, a novel automatic coronary artery segmentation method based on region growing with annular and spherical sector partition is proposed to improve search efficiency and quality of blood vessel segmentation. In our proposed method, the region is divided into a series of annular sectors based on the characteristics of vascular shape in the two-dimensional image, while the space is divided into spherical sectors according to the shape and tendency of the vessel in the three-dimensional image. The proposed method has been tested by 6 groups of data set, the efficiency and quality of the automatic segmentation has been significantly improved. Not only can the coronary artery and its adhesion tissue be successfully separated, but also the coronary arteries and their small branches can be detected. Furthermore, compared to the multiscale region growing method, our proposed method is able to search more branches. Also it is able to achieve higher total coverage ratio and Dice Similarity Coefficient ratio.

Effectively obtaining the centerline of the coronary arteries is very important to the clinical medical diagnosis. Cellular automata are extensively used to find fine solutions in many complex hard. However, it has the problem of large amount of computation. In this paper, we propose an automatic extraction method of coronary centerline based on cellular automata. Different from the existing methods, the distance from boundary field is used to extract the centerline of the coronary arteries roughly, which can place the candidate voxel at the center of the coronary arteries and reduce the size of the data. Furthermore, the centerlines of the coronary arteries are extracted accurately by using cellular automata to find the shortest path. The proposed method can extract the centerlines of each branch of the coronary arteries simultaneously from a single source to multiple destinations. Furthermore, compared to the method in Ref [10], our proposed method works well, especially in the case of large vessel curvature, the voxels at the extracted centerline are relatively centered. In the last, quantitative evaluation is conducted on the coronary arteries and the results show that the proposed method achieves a smaller deviation distance.

The K-mean algorithm is simple and efficient in vascular segmentation. But in a K-means clustering algorithm, the number of clustering centers must be determined in advance and the random allocation of clustering centers seriously affects the efficiency of the algorithm. In order to solve this problem, an adaptive clustering algorithm based on K-means with the Gaussian curve fitting histogram is proposed, which is used to segment circular-like blood vessels. Firstly, the histogram outline is fitted into the Gaussian curve by the least square method and the maximum value of the curve is obtained. The number of maximal values is counted as the number of clustering centers and the maximum points are used as the central points of clustering. Also, the K-means algorithm is improved by combining the intensity and position information of pixels. The proposed method has been tested by 6 groups of medical images. The algorithm has the advantages of less iterations, fast convergence speed and getting more than 95% Dice Similarity Measure value. The algorithm can effectively solve the determination of multi-seed points and vascular segmentation in medical image segmentation.