Component inspection device
A component inspection device including: a storage section configured to store at least a non-defective product images; a generation section configured to generate a defective product image using a machine learning model; a setting section configured for a user to set a parameter for component inspection; and an output section configured to perform inspection, by using the parameter, on the non-defective product image stored in the storage section and the defective product image generated by the generation section, the output section being configured to output an inspection result.
1 . A component inspection device comprising:
a storage section configured to store at least a non-defective product image;
a generation section configured to generate a defective product image using a machine learning model;
a setting section configured for a user to set a parameter for component inspection; and
an output section configured to perform inspection, by using the parameter, on the non-defective product image stored in the storage section and the defective product image generated by the generation section, the output section being configured to output an inspection result, wherein the output section outputs, as the inspection result, an inspection target image, information indicating whether the inspection target image is generated by the generation section or not, and information indicating which the inspection target image is determined to be, by using the parameter, from among a non-defective product or a defective product.
2 . The component inspection device according to claim 1 , wherein the generation section generates a defective product image by applying the machine learning model to at least any of the non-defective product images stored in the storage section.
3 . The component inspection device according to claim 1 , wherein the machine learning model is a deep generative model.
4 . The component inspection device according to claim 1 , further comprising:
a designation section configured for a user to designate a type of a defect of the defective product image to be generated by the generation section,
wherein the generation section generates the defective product image comprising a defect of the type designated by the designation section.
5 . The component inspection device according to claim 4 , wherein the designation section displays a designation screen configured for a user to designate a plurality of feature amounts representing defects.
6 . The component inspection device according to claim 5 , wherein the designation screen is configured for a user to designate a region in a feature space comprising a plurality of feature amounts.
7 . The component inspection device according to claim 6 , wherein an image representing a defect corresponding to the feature space is displayed in at least any of the regions in the feature space.
8 . The component inspection device according to claim 5 , wherein the designation screen is configured for a user to individually designate each of values of the plurality of feature amounts.
9 . The component inspection device according to claim 1 , further comprising a suggestion section configured to suggest a type of a defective product image to be generated.
10 . The component inspection device according to claim 9 , wherein the suggestion section generates, by the generation section, a defective product image for each of a plurality of types of defects,
performs a clustering process of classifying the generated defective product images into a plurality of clusters,
determines to which cluster each of the defective product images stored in the storage section belongs, and
suggests, as a type of defective product image to be generated, a defect corresponding to a cluster in which a number of existing defective product images is less than or equal to a predetermined number.
11 . A support method for supporting parameter setting for component inspection in a component inspection device, the support method comprising:
a generation step of generating a defective product image using a machine learning model with respect to a stored non-defective product image;
a setting step of receiving, from a user, setting of a parameter for component inspection; and
an output step of performing an inspection on the stored non-defective product image and the generated defective product image by using the parameter and outputting an inspection result, wherein the inspection result includes an inspection target image, information indicating whether the inspection target image is generated by the generation section or not, and information indicating which the inspection target image is determined to be, by using the parameter, from among a non-defective product or a defective product.
12 . A non-transitory computer readable medium storing a program configured to cause a computer to execute each step of the method according to claim 11 .