METHOD AND APPARATUS FOR WEAKLY-SUPERVISED LEARNING VIDEO CORPUS MOMENT RETRIEVAL

摘要

A method of operating a video section search apparatus operated by at least one processor, comprising the steps of receiving a video including a sentence and a frame described by the sentence, extracting a sentence feature vector and a video feature vector, and the sentence feature and generating a video candidate region composed of at least one frame related to the sentence, based on a similarity between the vector and the video feature vector.