Artificial intelligence deep-learning based video object recognition system and method

摘要

The present invention pre-learns a variety of images with deep learning of artificial intelligence, analyzes images in frame units for distributed video images by reflecting the pre-learning results, and optionally analyzes audio information of video images. Thus, after deriving the keywords in chronological order, the present invention relates to a system and method for recognizing artificial intelligence based on deep learning based on a method of comparing with the time-series keywords of an accumulated video script. The present invention is "a system implemented by artificial intelligence, a dataset learning unit that deep-learns multiple object images as keywords; A DB unit for extracting keywords from the video script, arranging and storing them in time series; A video object analysis unit for extracting a video object as a keyword and extracting a relationship between objects on the image as a keyword through frame-by-frame image analysis; And a comparison determination unit that compares keywords of the DB unit and the video analysis unit to determine similarity. It provides an artificial intelligence deep learning based video recognition system comprising a.