ABNORMALITY INSPECTION DEVICE AND ABNORMALITY INSPECTION PROCEDURES

Abstract

Abnormality inspection device (1) comprising the following: a picture acquisition unit (11) that acquires a target image in which a target object (3) to be inspected is captured; a learning outcome acquisition unit (123) that acquires a result of machine learning of a forward domain conversion of an image or a backward domain conversion of an image, performed using a normal image in which the target object (3) is captured in a normal state, as training data; a target image analysis unit (122) that obtains a domain-transformed image by sequentially performing a forward domain conversion and a backward domain conversion on the target image acquired by the image acquisition unit (11), using the machine learning output acquired by the learning outcome acquisition unit (123); and a determination unit (14) that determines whether or not an abnormality occurs in the target object (3) captured in the target image by comparing the target image acquired by the image acquisition unit (11) and the domain-transformed image acquired by the target image analysis unit (122); and wherein the target image analysis unit (122) receives a forward domain-transformed image by performing forward domain conversion on the target image acquired by the image acquisition unit (11), using a machine learning result performed using the normal image as training data, and wherein the abnormality inspection device further comprises an abnormality analysis unit (16) which performs analysis on the abnormality using the forward domain-converted image when the determination unit (14) determines that the abnormality occurs in the target object (3) captured in the determination target image acquired by the image acquisition unit (11).