Human body ultrasonic detection real-time guide strategy based on deep learning

Abstract

The invention discloses a human body ultrasonic detection real-time guide strategy based on deep learning. The human body ultrasonic detection real-time guide strategy comprises a demonstrator for human body ultrasonic detection, a skill evaluation strategy based on deep learning and a real-time adjustment strategy based on a sampling principle. The demonstrator is used for collecting a pressure signal and an attitude signal in the detection process; the skill evaluation strategy divides a human body ultrasonic image according to the effective degree of contained information, a training set is made and used for training a classification neural network, and therefore the neural network capable of judging whether the ultrasonic image meets the diagnosis requirement or not is obtained; according to the real-time adjustment strategy, on the basis of the skill evaluation strategy, multi-modal information collected by a demonstrator is added into a deep learning process, a multi-modal information fusion neural network is trained, a sampling set is made, and a sampling principle is combined to realize a function of guiding real-time adjustment of an ultrasonic probe. The strategy can help to complete medical ultrasonic detection, and greatly promotes the application of artificial intelligence in the field of ultrasonic detection.

Description

technical field [0001] The invention belongs to the field of data collection and deep learning, and relates to a data collection, skill evaluation and real-time adjustment strategy for human body ultrasonic detection, and specifically relates to a method of collecting human body ultrasonic detection data, using the data for deep learning and combining sampling principles to achieve Strategies for skills assessment and real-time adjustment. Background technique [0002] In recent years, deep learning, as a rapidly developing machine learning method, emphasizes learning from massive data to solve problems that traditional machine learning algorithms such as high dimensionality, complexity and high noise in big data are difficult to deal with. The application scenarios are gradually expanding, such as biometrics, smart driving, financial e-commerce, industrial manufacturing, etc. Medical ultrasound imaging has always been a research hotspot and difficulty in machine learning: ...

AI Technical Summary

Problems solved by technology

Commonly used image recognition methods include: statistical decision-making methods, structural pattern recognition methods, fuzzy pattern recognition methods, support vector machines and artificial neural networks. These methods can achieve a certain degree of ultrasound image recognition work, but also have obvious disadvantages: Considering the ultrasound image, but ignoring many information during the doctor's operation of the ultrasound probe, including signals such as probe pressure and posture

How to obtain the data of the doctor’s operation of the ultrasound probe, how to distinguish the complex ultrasound images according to the imaging effect, how to combine the operation data with the ultrasound image for deep learning, and how to further guide the adjustment of the probe, these problems lack a reference, Systematic solution

Images