Method for detecting abnormality, non-transitory computer-readable recording medium storing program for detecting abnormality, abnormality detection apparatus, server apparatus, and method for processing information

A non-transitory computer-readable recording medium storing a program for detecting an abnormality that causes a computer to function as:

• an obtainer that obtains a chest X-ray image;
• a detector that detects, in the obtained chest X-ray image, boundary lines of images of anatomical structures whose ranges of X-ray transmittances are different from one another using a model constructed through machine learning before the detection;
• a setter that sets a third lung area in the chest X-ray image including at least one of a first lung area where one or more lungs and a heart overlap or a second lung area where one of the lungs and a liver overlap on a basis of the detected boundary lines;
• a determiner that extracts a vascular index indicating at least one of thickness or density of at least one pulmonary blood vessel present in an area included in the third lung area and that determines whether the area included in the third lung area is in an abnormal state on a basis of the extracted vascular index and a reference index based on indices extracted, using a method used to extract the vascular index, in advance from an area in chest X-ray images in a normal state corresponding to the area included in the third lung area;
• and
• an output controller that outputs, if it is determined that the area included in the third lung area is in an abnormal state, information indicating a result of the determination made by the determiner.

An abnormality detection apparatus comprising:

• an obtainer that obtains a chest X-ray image;
• a detector that detects, in the obtained chest X-ray image, boundary lines of images of anatomical structures whose ranges of X-ray transmittances are different from one another using a model constructed through machine learning before the detection;
• a setter that sets a third lung area in the chest X-ray image including at least one of a first lung area where one or more lungs and a heart overlap or a second lung area where one of the lungs and a liver overlap on a basis of the detected boundary lines;
• a determiner that extracts a vascular index indicating at least one of thickness or density of at least one pulmonary blood vessel present in an area included in the third lung area and that determines whether the area included in the third lung area is in an abnormal state on a basis of the extracted vascular index and a reference index based on indices extracted, using a method used to extract the vascular index, in advance from an area in chest X-ray images in a normal state corresponding to the area included in the third lung area;
• and
• an output controller that outputs, if it is determined that the area included in the third lung area is in an abnormal state, information indicating a result of the determination made by the determiner.

A server apparatus comprising:

• an obtainer that obtains a chest X-ray image;
• a detector that detects, in the obtained chest X-ray image, boundary lines of images of anatomical structures whose ranges of X-ray transmittances are different from one another using a model constructed through machine learning before the detection;
• a setter that sets a third lung area in the chest X-ray image including at least one of a first lung area where one or more lungs and a heart overlap or a second lung area where one of the lungs and a liver overlap on a basis of the detected boundary lines;
• a determiner that extracts a vascular index indicating at least one of thickness or density of at least one pulmonary blood vessel present in an area included in the third lung area and that determines whether the area included in the third lung area is in an abnormal state on a basis of the extracted vascular index and a reference index based on indices extracted, using a method used to extract the vascular index, in advance from an area in chest X-ray images in a normal state corresponding to the area included in the third lung area;
• and
• an output controller that outputs, if it is determined that the area included in the third lung area is in an abnormal state, information indicating a result of the determination made by the determiner.

A method for detecting an abnormality, the method comprising:

• obtaining, using a computer, a chest X-ray image;
• determining, using the computer, whether an area included in a third lung area in the obtained chest X-ray image including a first lung area where one or more lungs and a heart overlap and/or a second lung area where one of the lungs and a liver overlap is in an abnormal state;
• and
• outputting, if determining that the area included in the third lung area is in an abnormal state, information indicating that the area included in the third lung area is abnormal using the computer.

A method for processing information, the method comprising:

• obtaining, using a computer, a chest X-ray image;
• detecting, in the obtained chest X-ray image using the computer and a model constructed through machine learning before the detecting, boundary lines of images of anatomical structures whose ranges of X-ray transmittances are different from one another;
• setting, using the computer, a third lung area in the chest X-ray image including at least one of a first lung area where one or more lungs and a heart overlap or a second lung area where one of the lungs and a liver overlap on a basis of the detected boundary lines;
• extracting, using the computer, a vascular index indicating at least one of thickness or density of at least one pulmonary blood vessel present in an area included in the third lung area;
• and
• outputting, using the computer, the extracted vascular index.