Method of detecting an abnormality and abnormality detection device
A method of detecting an abnormality includes: acquiring an image of an inspection target as a captured image; generating a restored image by inputting the captured image into a first learning model, wherein the first learning model has learned with a normal image, wherein the normal image obtained by capturing a normal inspection target; generating a difference image between the captured image and the restored image; generating a restored difference image by inputting the generated difference image into a second learning model, wherein the second learning model has learned with a normal difference image; and detecting abnormality of the inspection target using the difference image and the restored difference image.
1. A method of detecting an abnormality comprising:
acquiring an image of an inspection target as a captured image;
generating a restored image by inputting the captured image into a first learning model, wherein the first learning model has learned with a normal image, wherein the normal image obtained by capturing a normal inspection target;
generating a difference image between the captured image of the inspection target and the restored image;
generating a restored difference image by inputting the generated difference image into a second learning model, wherein the second learning model has learned with a normal difference image; and
detecting abnormality of the inspection target using the difference image and the restored difference image.
2. The method of detecting an abnormality according to claim 1 ,
wherein in the detecting the abnormality of the inspection target includes:
generating an abnormality determination image using a difference between the difference image and the restored difference image, and
determining the abnormality of the inspection target using the abnormality determination image.
3. The method of detecting an abnormality according to claim 1 ,
wherein the first learning model generates the restored image by removing a positional deviation of the inspection target in the captured image.
4. The method of detecting an abnormality according to claim 1 ,
wherein the first learning model is a learning model using a generative adversarial network.
5. The method of detecting an abnormality according to claim 1 ,
wherein the second learning model is a learning model using a generative adversarial network.
6. The method of detecting an abnormality according to claim 1 ,
wherein the first learning model and the second learning model are the same type of learning model.
7. An abnormality detection device, comprising:
an acquisition unit configured to acquire an image of an inspection target as a captured image;
a restored image generating unit configured to generate a restored image by inputting the captured image into a first learning model, wherein the first learning model has learned with a normal image, wherein the normal image obtained by capturing a normal inspection target;
a difference image generating unit configured to generate a difference image between the captured image of the inspection target and the restored image;
a restored difference image generating unit configured to generate a restored difference image by inputting the generated difference image into a second learning model, wherein the second learning model has learned with a normal difference image; and
an identification unit configured to detect abnormality of the inspection target using the difference image and the restored difference image.