APPARATUS AND METHOD FOR STUDYING PATTERN OF MOVING OBJECTS USING ADVERSARIAL DEEP GENERATIVE MODEL

摘要

The present invention provides an apparatus for learning a moving pattern of an object using a neural network generation model, which includes a moving pattern obtaining unit which tracks a feature point of the moving object in an input image, learns each moving pattern, and classifies the moving pattern according to a topic, and a moving pattern learning unit for each image which learns the moving pattern according to an inputted image using moving pattern (topic) information stored in the moving pattern obtaining unit. Accordingly, the present invention can instantaneously confirm which movement pattern is present by receiving an image of a corresponding surveillance area after learning the image using DBN and GAN by receiving a CCTV camera image of one days worth about the surveillance area as a learning image.