Multi-threshold Segmentation Algorithm For Medical Image Based On Hybrid Spatial Filter And Interval Iteration

Image processing technology is widely used in various fields of life,and image segmentation is a basic task of image processing.In the field of medical image processing,image segmentation technology provides the necessary basis for subsequent tasks such as image registration and image fusion and affects subsequent treatment.Since the development of medical image segmentation technology,many methods have been proposed,but there is still no general segmentation technology.The threshold segmentation method is widely used in the field of medical image segmentation because of its simple model and ideal segmentation effect.There are many regions of interest in medical images,and the single-threshold segmentation method can no longer meet the actual needs,so it needs to be extended to a multi-threshold segmentation algorithm.As the number of thresholds increases,the traditional threshold segmentation algorithm increases exponentially in computational time complexity,which cannot meet the real-time clinical needs.At present,the improvement of multi-threshold segmentation algorithm is mostly aimed at reducing the time complexity,often ignoring the importance of segmentation accuracy.As the number of thresholds increases,the traditional threshold segmentation algorithm increases exponentially in computational time complexity,which cannot meet the real-time clinical needs.At present,the improvement of multi-threshold segmentation algorithm is mostly aimed at reducing the time complexity,often ignoring the importance of segmentation accuracy.To solve the above problems and further improve the efficiency and accuracy of the medical image segmentation algorithm,this paper makes the following two parts of research:(1)This paper proposes a multi-threshold segmentation algorithm based on region growing and interval iteration.The region growing method treats all gray levels of the image as a small region,selects the main region and merges the region with reference to the idea of maximum inter-class variance,reduces the threshold and the number of regions in iterations,until the preset conditions are met,and outputs the initial Segmentation threshold.The iterative interval in the image is divided according to the initial segmentation threshold,the interval iteration strategy is applied in the iterative interval,and the single-threshold Otsu segmentation algorithm is applied to refine the classification of small objects,and the final segmentation threshold is output.The experimental results show that the algorithm can achieve satisfactory segmentation results with a very small-time cost,and the time consumption decreases with the increase of the number of thresholds,and can be stabilized at about 0.28 S,fully meeting the real-time requirements in clinical practice.(2)To overcome the influence of noise on segmentation results,this paper proposes a multi-threshold segmentation framework based on hybrid spatial filters and weighting strategies.The hybrid spatial filter is composed of Gaussian,median,mean,and mean-median filters.The original image is processed by the hybrid spatial filter to obtain the mixed layer of the original image.A multi-threshold segmentation algorithm is used for the original image and the mixed layer image to obtain two sets of segmentation thresholds,and a weighting strategy is used to balance the two sets of thresholds to obtain more accurate segmentation thresholds.The experimental results show that the hybrid spatial filter can smooth the noise while retaining the detailed information in the medical image,which can significantly improve the accuracy and stability of the segmentation algorithm,and the time cost is not increased much,and it can still meet the real-time requirements.