Real-time image generation method for super-resolution B ultrasonic image

Abstract

The invention discloses a real-time image generation method for a super-resolution B-mode ultrasound image, and aims to solve the problem that the definition in a B-mode ultrasound real-time image generation process needs to be improved. The method comprises the following steps: establishing an MSE-based loss function; preparing two losses for different stages: using cross entropy loss to performbinary classification on the gray level of the black and white image, initializing network internal characteristics and attention parameters, and using MSE as a loss function of a later stage refiningresult; constructing the neural network according to the core block of the deep convolutional neural network; using a block structure to perform stacking to obtain a deep convolutional network, and generating a static graph; using an Adam optimizer to initialize the cross entropy loss in the step b, stopping training when the loss rate is reduced to be nearly flat, changing the loss into MSE loss, and performing refining to generate enough reasoning results to improve the PSNR; and exporting the weight, and integrating and operating the weight in the medical equipment by using the static graph. According to the invention, single-frame input and single-frame output can be realized, and a result superior to a manual design image enhancement algorithm is obtained.

Description

technical field [0001] The invention relates to an image processing method of medical equipment, in particular to a real-time image generation method for super-resolution B-ultrasound images. Background technique [0002] B-ultrasound image is an image signal with a low signal-to-noise ratio. In order to achieve a higher signal-to-noise ratio, sharper, and more informative images, in addition to increasing the sampling rate of the device itself, it can also be processed by image processing methods. Performance improvements. Super-resolution can significantly reduce the A / D circuit performance and precision requirements of the device, thus relying on image processing to make up for the lack of hardware. Typical super-resolution methods are based on the integration of temporal information, that is, continuously changing B-Mode images are regarded as video streams, and clearer images are obtained through multi-frame synthesis technology. Based on this method, it will directly...

AI Technical Summary

Problems solved by technology

[0004] The present invention overcomes the problem that the clarity needs to be improved in the B-ultrasound real-time image generation process i

Images