ARTIFICIAL INTELLIGENCE (AI)-BASED OBJECT INSPECTION SYSTEM AND METHOD

摘要

Disclosed are an AI-based object inspection system and method for inspecting a defect of an object, such as a glass substrate or the like, by using learning. The AI-based object inspection system may include a region of interest (ROI) detecting unit for generating an ROI image by detecting an ROI from an input image that is an image of the object, a classifying unit for classifying the ROI image generated by the ROI detecting unit, on the basis of a first classification model, and a data learning unit for generating a model by learning an image provided from the classifying unit. Here, when a new image that does not belong to a classification standard is present as a result of classifying the ROI image on the basis of the first classification model, the classifying unit may provide the new image to the data learning unit, the data learning unit may generate a new second classification model by reflecting the provided new image and provide the second classification model to the classifying unit, and the classifying unit may classify the ROI image on the basis of the provided second classification model.