Method for detecting virological disorders
Embodiments may relate to techniques for image augmentation used for the detection and screening of respiratory diseases, such as COVID-19 pneumonia. For example, For example, in an embodiment a system for detecting medical conditions may comprise an image acquisition component adapted to acquire a plurality of images of persons, at least some of whom have a medical condition, an image augmentation component adapted to generate a plurality of additional images from the acquired plurality of images using a plurality of image augmentation methods, and a training component adapted to train a machine learning model using the acquired plurality of images and the generated plurality of additional images to form a trained machine learning mode, wherein when the medical condition is COVID-19, not generating the plurality of additional images so as to improve accuracy of the machine learning model in recognizing presence or absence of COVID-19 in images.
1 . A system for detecting medical conditions comprising:
an image acquisition component adapted to acquire a plurality of images of persons, at least some of whom have a medical condition;
an image augmentation component adapted to when the medical condition is not a virological disorder generate a plurality of additional images from the acquired plurality of images using a plurality of image augmentation methods and when the medical condition is a virological disorder, not generate the plurality of additional images so as to improve accuracy of a machine learning model in recognizing presence of a virological disorder in images; and
a training component adapted to train the machine learning model, when the medical condition is not a virological disorder using the acquired plurality of images and the generated plurality of additional images to form a trained machine learning model, and when the medical condition is a virological disorder using the acquired plurality of images to form the trained machine learning model.
2 . The system of claim 1 , wherein the plurality of image augmentation methods comprises at least some of: translation in x-axis, translation in y-axis, random shear in x-axis, random shear in y-axis, random rotation, horizontal reflection, vertical reflection, scaling in x-axis, scaling in y-axis, and any combination thereof.
3 . The system of claim 2 , wherein the acquired plurality of images comprises at least one of X-ray images, CT images, and MRI images.
4 . The system of claim 1 , wherein the virological disorder is COVID-19.
5 . A method for detecting medical conditions, implemented in a computer system comprising a processor, memory accessible by the processor, and program instructions and data stored in the memory to implement the method comprising:
acquiring a plurality of images of persons, at least some of whom have a medical condition;
when the medical condition is not a virological disorder:
generating a plurality of additional images from the acquired plurality of images using a plurality of image augmentation methods, and
training a machine learning model using the acquired plurality of images and the generated plurality of additional images to form a trained machine learning model; and
when the medical condition is a virological disorder:
not generating the plurality of additional images so as to improve accuracy of the machine learning model in recognizing presence of a virological disorder in images.
6 . The method of claim 5 , wherein the plurality of image augmentation methods comprises at least some of: translation in x-axis, translation in y-axis, random shear in x-axis, random shear in y-axis, random rotation, horizontal reflection, vertical reflection, scaling in x-axis, scaling in y-axis, and any combination thereof.
7 . The method of claim 6 , wherein the acquired plurality of images comprises at least one of X-ray images, CT images, and MRI images.
8 . The method of claim 5 , wherein the virological disorder is COVID-19.
9 . A non-transitory computer program product for detecting medical conditions, the computer program product comprising a non-transitory computer readable storage having program instructions embodied therewith, the program instructions executable by a computer, to cause the computer to perform a method comprising:
acquiring a plurality of images of persons, at least some of whom have a medical condition;
when the medical condition is not a virological disorder:
generating a plurality of additional images from the acquired plurality of images using a plurality of image augmentation methods, and
training a machine learning model using the acquired plurality of images and the generated plurality of additional images to form a trained machine learning model; and
wherein when the medical condition is a virological disorder:
not generating the plurality of additional images so as to improve accuracy of the machine learning model in recognizing presence of a virological disorder in images.
10 . The computer program product of claim 9 , wherein the plurality of image augmentation methods comprises at least some of: translation in x-axis, translation in y-axis, random shear in x-axis, random shear in y-axis, random rotation, horizontal reflection, vertical reflection, scaling in x-axis, scaling in y-axis, and any combination thereof.
11 . The computer program product of claim 10 , wherein the acquired plurality of images comprises at least one of X-ray images, CT images, and MRI images.
12 . The computer program product of claim 9 , wherein the virological disorder is COVID-19.