财务人员必备excel技巧.docVIP

Exposing Digital Forgeries in Video by Detecting Duplication Weihong Wang Department of Computer Science Dartmouth College Hanover,NH03755 whwang@cs.dartmouth.edu Hany Farid Department of Computer Science Dartmouth College Hanover,NH03755 farid@cs.dartmouth.edu ABSTRACT With the advent of high-quality digital video cameras and sophisticated video editing software,it is becoming increas-ingly easier to tamper with digital video.A common form of manipulation is to clone or duplicate frames or parts of a frame to remove people or objects from a video.We describe a computationally e?cient technique for detecting this form of tampering. Categories and Subject Descriptors I.4[Image Processing]:Miscellaneous General Terms Security Keywords Digital Tampering,Digital Forensics 1.INTRODUCTION In the popular movie Speed,the characters created a doc-tored video by simply duplicating a short sequence of frames, Figure1.In so doing,they were able to conceal activity on the bus.This common and relatively easy to perform ma-nipulation can be used to remove people or objects from a video sequence or simply remove an undesired event from a video.When done carefully,this form of tampering can be very di?cult to detect. Techniques for detecting image duplication have previ-ously been proposed[2,5].These techniques,however,are computationally too ine?cient to be applicable to a video sequence of even modest length.We describe two compu-tationally e?cient techniques for detecting duplication.In the?rst,we show how to detect duplicated frames,and in the second,we show how to detect duplicated regions across frames.We also describe how these techniques can be adapted to yield a more computationally e?cient image Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for pro?t or commercial advantage and that copies bear this notice and the full citation on the?rst page.To co