To provide a defect classification
system being simple in preparation and maintenance and capable of recognizing a cause of a defect. The present invention provides a defect classification
system 10 including imaging means 11 for causing an imaging device G to image paper P2 that has passed through a dry part D and acquiring image data obtained by the imaging, detection means 12 for detecting a defect in the paper P2 in the image data, extraction means 13 for extracting a feature amount of the defect, calculation means 14 for causing a classification model 20 in which a reference feature amount is previously set to calculate a
certainty factor in a defect cause item on the basis of the feature amount of the defect, and display means 15 for displaying the
certainty factor, in which the classification model 20 is caused to learn the reference feature amount using
machine learning from a relationship between respective feature amounts of defects previously stored and a plurality of defect cause items.