AN EFFECTIVE APPROACH FOR VIDEO COPY DETECTION USING SIFT FEATURES

Abstract

A SIFT features is an effective approach for video copy detection. To detect and describe local features in images scale-invariant feature transform is used. The purpose of video copy detection is to decide whether a video segment is a copy of video from train video database or not. SIFT image features provide a set of features of such as object scaling and rotation. We first use dual threshold method to segment the videos into segments with homogeneous content which helps to extract key frames from each segment. SIFT features are extracted from the key frames of the segments. SIFT features are very flexible to the effects of "noise" in the image. SIFT Applications include object recognition, robotic mapping and 3D modeling, navigation, image stitching, video tracking, gesture recognition, , individual identification of wildlife and match moving.