System and Method for recognizing art objects in video combining deep learning technology and image feature extraction technology

摘要

The present invention relates to a system and a method for recognizing an artwork in an image realized by a computer device and, more specifically, to a system and a method for collecting images to realize a position of an artwork in an image, annotating the image to store the same, learning the image by a deep learning algorithm of artificial intelligence, and extracting and storing image features of artworks accompanied with copyright so as to determine whether an artwork is included in a query image and whether copyright of another person is infringed by the included artwork. A system for recognizing an artwork in an image and realized by a computer device comprises: an image collection analysis unit for collecting images including artworks, and storing annotation indicating a position of an artwork in a collected image; an artwork detection unit based on artificial intelligence for deep-learning a position and annotation of an artwork in a collected image and detecting an artwork from a query image; an artwork collection analysis unit for collecting artwork images accompanied with copyright to normalize the size and store extracted image features; and a query image reading unit for extracting image features to compare the same with the image features stored in the artwork collection analysis unit, after separating a detected artwork image from the query image to normalize the size, and deriving a recognition result.