US 12,333,790 B2
Diagnostic assistance apparatus and model generation apparatus
Kouji Take, Kyoto (JP); and Kazuki Matsui, Kyoto (JP)
Assigned to MURATA MANUFACTURING CO., LTD., Kyoto (JP)
Filed by Murata Manufacturing Co., Ltd., Kyoto (JP)
Filed on Aug. 24, 2022, as Appl. No. 17/821,850.
Application 17/821,850 is a continuation of application No. PCT/JP2020/011839, filed on Mar. 17, 2020.
Prior Publication US 2023/0005251 A1, Jan. 5, 2023
Int. Cl. G06K 9/00 (2022.01); G06T 7/00 (2017.01); G06V 10/774 (2022.01); G16H 30/40 (2018.01)
CPC G06V 10/7747 (2022.01) [G06T 7/0014 (2013.01); G16H 30/40 (2018.01); G06T 2207/20081 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A diagnostic assistance apparatus comprising:
a data acquisition unit configured to acquire a target medical image of a body part of a target examinee;
at least one processor comprising:
a first classification model trained, by unsupervised learning using a plurality of first learning medical images of normal body parts, to provide an evaluation of a degree of normality of the body part in the acquired target medical image by one-class classification; and
a second classification model trained, by supervised learning using a plurality of learning data sets, each learning data set comprising a combination of a second learning medical image and a correct label indicating whether the body part captured in the second learning medical image is normal, to provide the evaluation of the degree of normality of the body part captured in the acquired target medical image, the second learning medical images of the plurality of learning data sets comprising a normal medical image capturing a normal body part and an abnormal medical image capturing an abnormal body part,
the at least one processor being configured to:
by feeding the acquired target medical image to the trained first classification model and performing a first operation of the trained first classification model, acquire as a first result the degree of normality evaluated by the one-class classification on the body part of the target examinee captured in the target medical image;
by feeding the acquired target medical image to the trained second classification model and performing a second operation of the trained second classification model, acquire as a second result the degree of normality evaluated on the body part of the target examinee captured in the target medical image; and
in accordance with the first result and the second result, provide a determination of whether the body part of the target examinee captured in the target medical image is normal; and
an output device configured to output the determination.