Detection And Combination Of Open Contours In Medical Images

The contour of the image is the dividing line between the object and the background in the image.It can best reflect the significant changes in the image attributes,based on which the image contour detection technology often plays an important role in the process of target extraction in practical application.This paper will focus on the application of image open contour detection and segmentation technology in the field of medical image processing.As for open fractures such as lung fissure,blood vessels,neurons,there are many surrounding tissues with changeable features,which is easy to cause problems including the ambiguity of the boundary and the complexity of the microstructure distribution between different tissues.At the same time,due to the impact of the noise,motion artifacts and partial volume effects in the process of image generation are likely to make the original image gray level distribution very uneven,the local feature blurred,and the micro part of some organizations break at the end.All of these make the problem of medical image contour detection become more complicated.In view of the above difficulties,the medical image open contour is taken as the research object in this paper.The complete segmentation result is obtained through linear tracking algorithm and contour combination algorithm for contour extraction and contour reconstruction respectively.It could provide an important reference for dividing functional areas and locating lesions.The main contents of this paper include:1.According to the characteristics of blood vessels,lung fissure and other medical image open contour of the valley-shaped cross section,this paper presents a linear tracking method to realize the automatic segmentation of the image open contour.First,define the initial seed points according to the selected area of the image histogram,and carry out the process of linear tracking initialization.The shape of the cross section is used as the basis to calculate the cross-sectional profile parameter.It will determined the new linear tracking point,iterate until all initial seed points are traversed,and we can get the contour of the initial estimated figure.Finally,the method of map quantization and morphological post-processing is used to denoise and obtain the final contour detection results.2.According to the characteristics of the fast marching algorithm and the shortest path,a contour combination algorithm is proposed to realize the reconstruction of fracture profile in this paper.First,we should take partition processing to the fracture profile,assign different regions to different subscript and make all of its points as seed points.According to the predetermined speed function,perform fast marching algorithm evoluting from all the seed point.In the process of fast marching,when different regions evolve forward and collide,the first collision point would be marked as a saddle point and the subscript of the two regions should be recorded at the same time.Therefore,we can obtained a minimal action map and all the saddle points from the connected components when the fast marching algorithm is executed.Finally,according the minimal action map backpropagating from the saddle point to the two connected areas we can obtain the complete image contour.3.In order to verify the validity of the algorithm,we establish a comprehensive evaluation system in this paper.By comparing quantitatively and qualitatively the"gold standard" with the segmentation results,this system calculates the TPR,FDR,detection accuracy and the average running time.The comparison between experimental results and algorithm shows that the algorithm presented in this paper has good performance in segmentation,and anti noise.It can meet the requirements of clinical application.