Ultrasonic image acquisition intelligent positioning method and system based on deep learning network

Abstract

The invention discloses an ultrasonic image acquisition intelligent positioning method and system based on a deep learning network, and the method comprises the steps: scanning and obtaining a two-dimensional image and spatial position coordinates of the two-dimensional image, and storing the interested two-dimensional image for building a reference image library; obtaining three-dimensional volume data corresponding to the two-dimensional image, obtaining cuboid three-dimensional volume data corresponding to the three-dimensional volume data, and establishing a standard cuboid three-dimensional volume data space coordinate system; constructing a deep learning network, generating any two-dimensional section image and spatial position coordinates thereof by utilizing standard cuboid three-dimensional volume data, and training the deep learning network to meet the following conditions: when a two-dimensional scanning image is input, automatically outputting the spatial position coordinates of a virtual probe corresponding to the scanning image. The spatial position coordinates of the two-dimensional image section can be obtained in real time, the probe is prompted in real time to scan the position of the section in the human body, a doctor is guided to find the interested or marked section, time is saved, and the diagnosis efficiency is improved.

Description

technical field [0001] The present invention relates to the technical field of ultrasonic imaging, in particular to an intelligent positioning method and system for ultrasonic imaging based on a deep learning network. Background technique [0002] Medical ultrasound imaging systems are suitable for clinical diagnosis of organs and tissues such as liver, kidney, and breast. How to establish the mapping relationship between ultrasound 2D slices (or 2D scan images) and 3D volume data. In the existing medical ultrasound system, how to map the 2D slice (or 2D scan image) to the corresponding 3D volume data is relatively mature, and the commonly used ones are similar: add a magnetic field tracking sensor to the probe to record the geometric position of the 2D slice, use Image frame-to-frame correlation during probe movement, use of 3D volumetric probes, and ABUS (Automated Breast Ultrasound System) that uses robotic arms to move ultrasound probes. However, in terms of mapping 3D...

AI Technical Summary

Problems solved by technology

[0003] To sum up, in the existing clinical ultrasonography, the image quality and diagnostic accuracy of ordinary hand-held probes depend on the experience and skill of the operating physician, so the optimal slice positions drawn by doctors of different levels may deviate

Images