Date of Award

Spring 2013

Document Type

Thesis

Degree Name

Master of Science in Electrical Engineering - (M.S.)

Department

Electrical and Computer Engineering

First Advisor

Yun Q. Shi

Second Advisor

Edwin Hou

Third Advisor

Tan-Feng Sun

Abstract

Recently, due to availability of inexpensive and easily-operable multimedia tools, digital multimedia technology has experienced drastic advancements. At the same time, video forgery becomes much easier and makes more difficult to validate the video content. Consequently, the origin and integrity of video can no longer be taken for granted. A methodology is developed that is capable of detecting the video frame insertion based on the constraint of human visual perception. The main idea is based on the so-called differential sensitivity. That is, that the variation of brightness of neighboring video frames has some constraint. First, the video sequence is partitioned into short and overlapping sub-sequences. Second, the ratio of the temporal variation of brightness calculated at the beginning and the ending frames of each sub-sequence is computed and compared with a threshold to determine the approximate location of the video frame insertion. Third, a procedure is conducted to determine the exact location of the insertion. The success of simulation works on more than 200 video sequences. The precision rate of detection is about 94.09%, and the precision rate of detecting location of frame insertion is 84.88% on testing database

Share

COinS