A method and system for detecting a near-duplicate of a video clip, according to the present invention, comprise a server including: a first codebook extraction module for extracting at least one first image frame and at least one first feature vector corresponding to each of the extracted image frames, from a reference image including at least one image frame, and for grouping the at least one first feature vector and generating at least one first center vector of the grouped at least one first feature vector in a first vector space in which the at least one first feature vector exists; a first codebook database in which the first center vectors is stored; a first feature extraction module for extracting at least one second image frame for a first time period, and extracting at least one second feature vector corresponding to each of the extracted image frames, from the reference image; and a first vector quantization module for calculating the distance between each of the second feature vectors and the first center vector to give each second feature vector a weighting, and generating a first feature vector histogram in which the second feature vectors each given the weighting are accumulated.
展开▼
