LEARNING METHOD AND LEARNING DEVICE USING MULTIPLE LABELED DATABASES WITH DIFFERENT LABEL SETS AND TESTING METHOD AND TESTING DEVICE USING THE SAME

摘要

The present invention relates to a method for learning a CNN using a plurality of labeled databases having different label sets, wherein the learning apparatus is annotated with a class in which each object belongs to a corresponding category and a class corresponding to the object. constructing a learning database including image data for each category and a GT label set for each category when corresponding to each of the attached information, and the GT label set corresponding to the image data set; (b) Receive a specific image belonging to a specific image data set corresponding to a specific class in the learning database as an input image, generate a feature map, and classify by category corresponding to a specific object included in the input image based on the feature map generating a resu (c) learning the parameters of the CNN using the loss for each category; a method including a method is provided.