Data processing device and data processing method

Abstract

This invention provides a data processing apparatus that, when multiple learning models exist, can generate learning models that both save operator time and achieve desired target performance. Specifically, the image processing apparatus 100 includes: a processing unit 10 that performs: a selection process, which selects a learning model M from a plurality of learning models M; a learning model process, which includes a learning process of learning the selected learning model M and an input / output process of receiving output from the selected learning model M; and a determination process, which determines whether a detection result based on the presence or absence of anomalies in the output of a learning model M constitutes a preset target performance. The processing unit 10 is configured to repeat the selection process, the learning model process, and the determination process until the detection result in the determination process constitutes the target performance, and then designates the learning model M that achieves the target performance as the qualified learning model.

Images