Respiratory signal extraction method and device for angiography image sequence

Abstract

The embodiment of the invention provides a respiratory signal extraction method and device for an angiography image sequence, and the method comprises the steps: obtaining an angiography image sequence, inputting the angiography image sequence into a pre-trained neural network, and outputting a respiratory signal sequence; wherein the neural network comprises a self-encoder layer which is used forinputting an angiography image sequence and outputting background images of each frame of image in the angiography image sequence to form a background image sequence; an optical flow network layer used for inputting the background image sequence and outputting motion information of the background image sequence; and a recurrent neural network layer used for inputting the motion information of thebackground image sequence and outputting the respiratory signal sequence. According to the embodiment of the invention, the respiratory signal sequence can be automatically and accurately extracted.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of image processing, and more specifically, to a method and device for extracting respiratory signals of a contrast image sequence. Background technique [0002] Angiography is an interventional detection method in which a contrast agent is injected into blood vessels. Because X-rays cannot penetrate the contrast agent, the image displayed by the contrast agent under X-ray is considered to be a blood vessel. Currently, contrast images are the most common imaging modality in thoracic and abdominal vascular interventional procedures. In thoracic and abdominal related operations, the patient's breathing will cause the displacement of various tissues and organs, so in the image, the position of the contrast agent will also change, which will bring difficulties to the doctor's positioning operation. In order to compensate the displacement of tissues and organs in the image with respiration,...

AI Technical Summary

Problems solved by technology

They all have certain limitations: In the first category of methods, the region of interest is the diaphragm structure

The limitation of this method is that the specific dimensionality reduction modeling process needs to input a sequence of images first, which is not conducive to real-time operation during surgery

Images