Breast lump ultrasonic image ultrasonic region separation method

Abstract

The invention discloses a breast lump ultrasonic image ultrasonic region separation method, which comprises the following steps of: 1.1, filtering a breast lump ultrasonic image I by using a pyramid mean shift algorithm to obtain a filtered image If; 1.2, segmenting the image If by using a watershed algorithm; 1.3, calculating the minimum gray scale minGray in the specific region of interest in the Is; 1.4, traversing Is, for pixels with the gray level equal to minGray, assigning the gray level of the pixels as a foreground, namely a target lump area, otherwise, assigning the gray level of thepixels as a background, and obtaining a final lump segmentation result binary image. According to the method, the automation level is high when the breast lump ultrasonic image is segmented, completeand sufficient automatic segmentation is achieved, and manual interaction is avoided.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a method for automatic segmentation of medical ultrasound images, which is a method for separating ultrasound regions of ultrasound images of breast masses. Background technique [0002] Breast lumps are one of the most common diseases among women. At present, the most effective detection and diagnosis method is mammography. However, the low specificity of mammography has resulted in a large number of unnecessary biopsies, which not only brings pain to patients, but also increases costs. In addition, ionizing radiation from mammography poses health risks to patients and physicians. Ultrasound imaging has the characteristics of low cost, non-invasiveness, and real-time performance, and has become one of the important means of breast mass detection. However, breast ultrasound images have low contrast, speckle noise, and tissue-related texture structures, which will bring...

AI Technical Summary

Problems solved by technology

These methods have some major disadvantages: (1) most methods require manual interaction, such as manual selection of seed points or initial contours; (2) most methods are sensitive to speckle noise, and because ultrasound images have low contrast and tissue-related textures, so Accurate segmentation is difficult; (3) Most methods have high algorithm complexity and long processing time, which are difficult to meet clinical requirements

However, these methods have the following problems: (1) The algorithms for automatically finding seed points or regions of interest are often not robust, and are only effective for some images, which affects the accuracy of automatic segmentation; (2) The algorithm complexity is high, and processing Time-consuming, it is difficult to meet the clinical speed requirements for automatic segmentation

The watershed algorithm is a fast segmentation method widely used in image segmentation, but due to the inherent characteristics of ultrasound images, too many regions will be generated during the submersion process, which will lead to slow segmentation and low accuracy.

Images