Real-time tracking and quantitative analysis method based on liver ultrasound contrast image

Abstract

The invention belongs to the technical field of computer vision and digital image processing, and specifically relates to a real-time tracking and quantitative analysis method based on a liver ultrasound contrast image. The method mainly comprises the following steps: 1) selecting a region of interest from an initial frame of a liver sequence image, and determining a candidate region through combination of an optical flow method and a particle filtering method; 2) extracting features of a candidate sample by utilizing a deep neural network, and obtaining a tracking result; 3) judging accuracy of the current tracking result through similarity match; and 4) quantifying radiography parameters of an observation point area, eliminating fluctuation interference in the image, and fitting a time intensity curve through a cubic spline interpolation method to allow doctor's vision and differential diagnosis to be more visual. Compared with the prior art, the method, by utilizing the feature that tissue deformation has period, and through a global and local area combination method, has dominant advantages in tracking effect and speed.

Description

technical field [0001] The invention belongs to the technical fields of computer vision and digital image processing, and in particular relates to a real-time tracking and quantitative analysis method based on liver ultrasound contrast images. Background technique [0002] At present, contrast imaging has gradually become an important means of medical diagnosis. The tracking and quantitative analysis of observation points in the region of interest in the contrast image sequence can assist clinicians to make scientific judgments and decisions on liver diseases, save doctors' diagnosis time, and reduce the impact of subjective judgments. [0003] With the development of contrast imaging technology and tracking methods, the objective quantitative analysis of the region of interest has also gradually developed. In reality, the qualitative quantification of contrast parameters is to track and analyze the selected observation points of the region of interest. The time intensity cu...

AI Technical Summary

Problems solved by technology

[0007] In view of the deficiencies in the prior art, the present invention provides a method for tracking and quantitatively analyzing the region of interest in liver contrast images. Aiming at the problem that the observation point drifts or even disappears due to tissue deformation during the tracking of key regions in the contrast images, a deep learning model and traditional method are adopted. The method of combining algorithms tracks the region of interest, and then uses the interpolation method to accurately quantitatively analyze the time-intensity curve

Images