STREAMING SERVER AND METHOD FOR OBJECT PROCESSING IN MULTI-VIEW VIDEO USING THE SAME

摘要

The present invention relates to a streaming server and an object processing method in a multi-view video using the same, in particular, in the object processing method in a multi-view video performed by a streaming server, a) one arranged omnidirectionally from a content production terminal Receiving image data for each viewpoint captured using the above cameras in real time; b) performing an image correction algorithm for performing image stitching and color tone correction on the image data for each viewpoint, and merging them into one multi-view image; c) In the merged multi-view image, a region of interest is set based on the viewing angle of the user, and at least one object image is extracted from the region of interest using an object detection model based on artificial intelligence, and then meta for the extracted object image. Storing the data; d) generating an augmented reality (AR) content corresponding to the object image, and integrating the object image, AR content, and metadata into the multi-view image to generate and store a 360-degree streaming image; And e) providing the 360-degree streaming image to the user terminal at the request of the user terminal, and when an object of interest is selected by the user terminal while the 360-degree streaming image is being played, AR content included in the object of interest. Including the step of real-time outputting, wherein the artificial intelligence-based object detection model collects a plurality of 360-degree streaming images through a network and stores them as a training image, and performs a labeling operation on the object image among the training images. Thus, the labeling data is stored in a database, and the labeling data is learned based on the training image stored in the database.