Liver vessel ultrasound image target identification and tracking method based on improved U-net network and LSTM network

Abstract

The invention discloses a liver vessel ultrasound image target identification and tracking method based on an improved U-net network and an LSTM network. The method comprises: 1, preprocessing an ultrasonic image sequence; 2, training an ROI extraction model, and segmenting an area from the ultrasonic image in the step 1; 3, realizing accurate segmentation of the target in the region in the step 2 based on the improved U-net network; 4, classifying the ultrasonic image sequence in the step 1 by using a CNN-LSTM network; and 5, based on a segmentation result in the step 3 and a classification result in the step 4, realizing accurate prediction of a target position in the ultrasonic image sequence by using an LSTM network. The invention aims to solve the problem that the existing method cannot perform target identification and real-time tracking on a dynamic ultrasonic image sequence at the same time.

Description

technical field [0001] The invention relates to the technical field of image processing and the field of medical aided diagnosis, in particular to a target recognition and tracking method of liver blood vessel ultrasound images based on improved U-net network and LSTM network. Background technique [0002] Due to the real-time characteristics of medical ultrasound imaging, this technology is widely used in the identification and tracking of lesions in liver tumor ablation and minimally invasive treatment. However, respiratory movement will change the position of the lesion and obstacle, which increases the difficulty of accurate obstacle avoidance and lesion location. [0003] Target recognition and tracking based on ultrasound images mainly include: 1) target recognition in ultrasound images; 2) target tracking in ultrasound images. The target recognition of ultrasound images is to accurately draw the contours of valuable targets such as blood vessels and tumors in the ima...

AI Technical Summary

Problems solved by technology

[0007] The present invention provides a liver vessel ultrasound image target recognition and tracking method based on the improved U-net network and LSTM network, in order to solve the problem that existing methods cannot simultaneously perform target recognition and real-time tracking on dynamic ultrasound image sequences

Images