Follicle information extraction method based on ultrasonic image and related device

Abstract

The invention relates to the technical field of ultrasound, in particular to a follicle information extraction method based on an ultrasound image and a related device, which are used for solving the problems of how to improve the accuracy of identifying a plurality of follicles in the ultrasound image and how to improve the generalization degree of a follicle information extraction model. According to the method and the device, the first image is obtained by performing histogram equalization on the ultrasonic image, so that the follicle edge is displayed more clearly. Inputting the first image into an encoder module to obtain a first feature, and then obtaining a second feature through a multi-scale multi-path fusion module; inputting the second feature into a shape flow module to obtain a third feature, so that the follicle edge is clearer; and inputting the third feature and the second feature into a decoder module to obtain a mask image, and extracting and displaying the number of the follicles and / or the size information of the follicles based on the mask image. Therefore, the follicle recognition accuracy is improved, and the multi-scale multi-path fusion model improves the fusion capability of high-dimensional and low-dimensional features, so that the generalization degree of the follicle information extraction model is also improved.

Description

technical field [0001] The present application relates to the field of ultrasound technology, in particular to a method for extracting follicle information based on ultrasound images and a related device. Background technique [0002] When processing the ultrasound images of follicles, the obtained ultrasound images are blurry and noisy, and the structure of the ovary is complex, and the size and number of follicles are uneven, which makes it difficult to generalize the identification model of follicles. In ultrasound images, there are other hypoechoic areas besides follicles, such as blood vessels, cysts, etc., which may easily cause misidentification. Although in the related art, the deep learning algorithm based on the convolutional neural network improves the accuracy of the segmentation, it has a high false detection rate for multiple and messy follicle distribution. Therefore, how to improve the accuracy of identifying multiple follicles in ultrasound images and impro...

AI Technical Summary

Problems solved by technology

[0003] The present application discloses a method for extracting follicle information based on ultrasound images and related devices, which are used to solve the problem of how to improve the accuracy of identifying multiple follicles in ultrasound images and improve the generalization degree of the follicle information extraction model

Images