US 12,364,452 B2
Diagnostic information processing method and apparatus based on medical image, and storage medium
Lei Shi, Hangzhou (CN); Xuan Zang, Hangzhou (CN); and Jing Shi, Hangzhou (CN)
Appl. No. 17/760,185
Filed by HANGZHOU YITU HEALTHCARE TECHNOLOGY CO., LTD., Hangzhou (CN)
PCT Filed Feb. 5, 2021, PCT No. PCT/CN2021/075379
§ 371(c)(1), (2) Date Aug. 4, 2022,
PCT Pub. No. WO2021/155829, PCT Pub. Date Aug. 12, 2021.
Claims priority of application No. 202010081111.5 (CN), filed on Feb. 5, 2020; application No. 202010083597.6 (CN), filed on Feb. 7, 2020; and application No. 202010096657.8 (CN), filed on Feb. 17, 2020.
Prior Publication US 2023/0070249 A1, Mar. 9, 2023
Int. Cl. A61B 6/00 (2024.01); A61B 6/03 (2006.01); G06T 7/00 (2017.01); G06T 7/11 (2017.01); G06T 7/62 (2017.01); G06T 11/00 (2006.01); G06T 11/20 (2006.01); G16H 30/40 (2018.01); G16H 50/30 (2018.01)
CPC A61B 6/5217 (2013.01) [A61B 6/032 (2013.01); G06T 7/0016 (2013.01); G06T 7/11 (2017.01); G06T 7/62 (2017.01); G06T 11/008 (2013.01); G06T 11/206 (2013.01); G16H 30/40 (2018.01); G16H 50/30 (2018.01); G06T 2207/10081 (2013.01); G06T 2207/20021 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30061 (2013.01)] 9 Claims
OG exemplary drawing
 
1. A diagnostic information processing method based on a medical image, comprising the steps of:
acquiring a first lung medical image of an examined object;
acquiring image parameters of an affected part in the first lung medical image; which comprises the step of: inputting at least one first lung medical image into a neuron network to determine a volume of the affected part in the first lung medical image; and
determining, according to the image parameters of the affected part, a disease grade of lungs of the examined object corresponding to information of the first lung medical image,
wherein the neuron network comprises:
a first detection model configured to detect a candidate patch shadow, a cutting model, a second detection model configured to detect a patch shadow interval, and a volume calculation model configured to calculate the volume of the affected part; and
the step of inputting at least one first lung medical image into a neuron network to determine a volume of the affected part in the first lung medical image comprises the steps of:
passing the at least one first lung medical image through N consecutive convolution feature extraction modules in the first detection model, so that the N consecutive convolution feature extraction modules obtain image features of a patch shadow in the first lung medical image, wherein N is a positive integer;
inputting image features of the affected part in the first lung medical image into a fully connected layer in the first detection model, so that the fully connected layer outputs the candidate patch shadow based on the image features;
passing the candidate patch shadow through the cutting model, so that the cutting model cuts the candidate patch shadow in different directions in space multiple times to obtain multiple section images of the candidate patch shadow in multiple directions in space;
passing multiple consecutive section images through M consecutive convolution feature extraction modules in the second detection model, so that the M consecutive convolution feature extraction modules obtain image features of the section images, wherein M is a positive integer;
inputting the image features of the section images into a fully connected layer in the second detection model, so that the fully connected layer outputs patch shadow information based on the image features; and
passing the patch shadow information through the volume calculation model, so that the volume calculation model calculates the volume of the affected part in the first lung medical image.