Performance Evaluation of Digital Video Watermarking on GPU Parallel Computing

碩士 === 東海大學 === 資訊工程學系 === 102 === Along with the scientific and technological progress in the information explosion era, human life style has become much more technological and increasingly integrated with computer usages. With the rapid development of the Internet technology, digital archiving has...

Full description

Bibliographic Details
Main Authors: JHANG,HONG-YOU, 張紘侑
Other Authors: Chu-Hsing Lin
Format: Others
Language:zh-TW
Published: 2014
Online Access:http://ndltd.ncl.edu.tw/handle/64414965318754626402
Description
Summary:碩士 === 東海大學 === 資訊工程學系 === 102 === Along with the scientific and technological progress in the information explosion era, human life style has become much more technological and increasingly integrated with computer usages. With the rapid development of the Internet technology, digital archiving has become the trend of the past few years, and transmission of multimedia data over the Internet has become simple and easy; as a result, the confidentiality of information has become increasingly important. Because of advances in technology, digital means used in data theft, embezzlement and the like are also developing; hence, important personal information or confidential data, neglected and unprotected, are vulnerable to theft and improper use. Therefore, the confidentiality of information has become increasingly important, and study of intellectual property rights of digital videos has turned into a hot issue in the digital age nowadays. In this thesis, we proposed a technique that focuses on combining Graphic Processing Unit (GPU), CUDA, and the Distributed Discrete Wavelet Transform, which is a robust watermarking algorithm previously proposed by us. Finally, we conduct performance comparison of embedding watermarks on videos using GPU or CPU to investigate the effectiveness of these two different computing approaches. We have obtained some encouraging research results from the experiments: much computation time is saved by using GPU in aid of operations instead of using CPU alone. Although the results are still not ideal in performance, but the robustness of watermarking is quite satisfying.