Computer system, method, and program for diagnosing subject
The present invention is to provide a computer system, a method, and a program for diagnosing a subject that improve the accuracy of diagnosis by combining a plurality of time-series image data more than that by a conventional single image analysis. The computer system for diagnosing a subject acquires a plurality of first subject images with time series variation of the subject, analyzes the acquired first subject images, acquires a plurality of second subject images with time series variation of another subject in the past, analyzes the acquired second subject images, checks the analysis result of the first subject images and the analysis result of the second subject images, and diagnoses the subject based on the check result.
1. A computer system, comprising:
one or more processors that:
acquire a plurality of first subject images with time series variation of a subject;
analyze each of colored certain parts included in each of the acquired first subject images, each of the colored certain parts in each of the acquired first subject images being marked in a different color, and a same certain part of the acquired first subject images being marked in a same color;
acquire a plurality of second subject images with time series variation of another subject in the past;
analyze each of colored certain parts included in each of the acquired second subject images, each of the colored certain parts in each of the acquired second subject images being marked in a different color, and a same certain part of the acquired second subject images being marked in a same color;
check the analysis result of the first subject images and the analysis result of the second subject images; and
diagnose the subject based on the check result,
wherein the one or more processors calculate a degree of similarity between the first subject images and the second subject images based on the check result and diagnoses the subject, and
wherein the one or more processors, when the degree of similarity is less than a predetermined value,
acquire a plurality of third subject images with time series variation of yet another subject in the past,
analyze each of colored certain parts included in each of the acquired third subject images, each of the colored certain parts in each of the acquired third subject images being marked in a different color, and a same certain part of the acquired third subject images being marked in a same color,
further check the analysis result of the first subject images and the analysis result of the third subject images, and
diagnose the subject based on the further check result.
2. The computer system according to claim 1 , wherein the one or more processors check a feature point of each of the colored certain parts in the first subject images and a feature point of each of the colored certain parts in the second subject images.
3. The computer system according to claim 1 , wherein the one or more processors check a feature amount of each of the colored certain parts in the first subject images and a feature amount of each of the colored certain parts in the second subject images.
4. The computer system according to claim 1 , wherein the one or more processors diagnose a risk of acquiring a disease for the subject based on the check result.
5. The computer system according to claim 1 , wherein the one or more processors perform machine-learning on a feature point of the second subject images and analyzes the acquired first subject images.
6. The computer system according to claim 1 , wherein the one or more processors perform machine-learning on a feature amount of the second subject images and analyzes the acquired first subject images.
7. The computer system according to claim 1 , wherein the first subject images are eye-fundus images, the one or more processors analyze the acquired eye-fundus images which are marked, and diagnose the subject's glaucoma.
8. A method for diagnosing a subject by a computer system, comprising:
acquiring a plurality of first subject images with time series variation of the subject;
analyzing each of colored certain parts included in each of the acquired first subject images, each of the colored certain parts in each of the acquired first subject images being marked in a different color, and a same certain part of the acquired first subject images being marked in a same color;
acquiring a plurality of second subject images with time series variation of another subject in the past;
analyzing each of colored certain parts included in each of the acquired second subject images, each of the colored certain parts in each of the acquired second subject images being marked in a different color, and a same certain part of the acquired second subject images being marked in a same color;
checking the analysis result of the first subject images and the analysis result of the second subject images; and
diagnosing the subject based on the check result,
the method further comprising:
calculating a degree of similarity between the first subject images and the second subject images based on the check result; and
when the degree of similarity is less than a predetermined value,
acquiring a plurality of third subject images with time series variation of yet another subject in the past,
analyzing each of colored certain parts included in each of the acquired third subject images, each of the colored certain parts in each of the acquired third subject images being marked in a different color, and a same certain part of the acquired third subject images being marked in a same color,
further checking the analysis result of the first subject images and the analysis result of the third subject images, and
diagnosing the subject based on the further check result.
9. A non-transitory computer-readable medium comprising instructions that when executed by a computer system cause the computer system to execute
acquiring a plurality of first subject images with time series variation of the subject;
analyzing each of colored certain parts included in each of the acquired first subject images, each of the colored certain parts in each of the acquired first subject images being marked in a different color, and a same certain part of the acquired first subject images being marked in a same color;
acquiring a plurality of second subject images with time series variation of another subject in the past;
analyzing each of colored certain parts included in each of the acquired second subject images, each of the colored certain parts in each of the acquired second subject images being marked in a different color, and a same certain part of the acquired second subject images being marked in a same color;
checking the analysis result of the first subject images and the analysis result of the second subject images; and
diagnosing the subject based on the check result,
wherein the instructions cause the computer system to further execute:
calculating a degree of similarity between the first subject images and the second subject images based on the check result; and
when the degree of similarity is less than a predetermined value,
acquiring a plurality of third subject images with time series variation of yet another subject in the past,
analyzing each of colored certain parts included in each of the acquired third subject images, each of the colored certain parts in each of the acquired third subject images being marked in a different color, and a same certain part of the acquired third subject images being marked in a same color,
further checking the analysis result of the first subject images and the analysis result of the third subject images, and
diagnosing the subject based on the further check result.