東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Perceptually based methods for robus...

Monga, Vishal.

FindBook

Google Book

Amazon

博客來

Perceptually based methods for robust image hashing.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Perceptually based methods for robust image hashing./
作者:	Monga, Vishal.
面頁冊數:	120 p.
附註:	Source: Dissertation Abstracts International, Volume: 67-01, Section: B, page: 0445.
Contained By:	Dissertation Abstracts International67-01B.
標題:	Engineering, Electronics and Electrical. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3204165
ISBN:	9780542500022

Perceptually based methods for robust image hashing.
Monga, Vishal.

Perceptually based methods for robust image hashing. - 120 p.

Source: Dissertation Abstracts International, Volume: 67-01, Section: B, page: 0445.

Thesis (Ph.D.)--The University of Texas at Austin, 2005.

Hash functions are frequently called message digest functions. Their purpose is to extract a short binary string from a large digital message. A key feature of conventional cryptographic (and other) hashing algorithms such as message digest 5 (MD5) and secure hash algorithm 1 (SHA-1) is that they are extremely sensitive to the message; i.e., changing even one bit of the input message will change the output dramatically. However, multimedia data such as digital images undergo various manipulations such as compression and enhancement. An image hash function should instead take into account the changes in the visual domain and produce hash values based on the image's visual appearance. Such a function would facilitate comparisons and searches in large image databases. Other applications of a perceptual hash lie in content authentication and watermarking.

ISBN: 9780542500022Subjects--Topical Terms:

626636
Engineering, Electronics and Electrical.

Perceptually based methods for robust image hashing.
LDR:03525nmm 2200301 4500 001 1822184
005 20061127121654.5
008 130610s2005 eng d
020 $a 9780542500022
035 $a (UnM)AAI3204165
035 $a AAI3204165
040 $a UnM $c UnM
100 1 $a Monga, Vishal. $3 1911335
245 1 0 $a Perceptually based methods for robust image hashing.
300 $a 120 p.
500 $a Source: Dissertation Abstracts International, Volume: 67-01, Section: B, page: 0445.
500 $a Adviser: Brian L. Evans.
502 $a Thesis (Ph.D.)--The University of Texas at Austin, 2005.
520 $a Hash functions are frequently called message digest functions. Their purpose is to extract a short binary string from a large digital message. A key feature of conventional cryptographic (and other) hashing algorithms such as message digest 5 (MD5) and secure hash algorithm 1 (SHA-1) is that they are extremely sensitive to the message; i.e., changing even one bit of the input message will change the output dramatically. However, multimedia data such as digital images undergo various manipulations such as compression and enhancement. An image hash function should instead take into account the changes in the visual domain and produce hash values based on the image's visual appearance. Such a function would facilitate comparisons and searches in large image databases. Other applications of a perceptual hash lie in content authentication and watermarking.
520 $a This dissertation proposes a unifying framework for multimedia signal hashing. The problem of media hashing is divided into two stages. The first stage extracts media-dependent intermediate features that are robust under incidental modifications while being different for perceptually distinct media with high probability. The second stage performs a media-independent clustering of these features to produce a final hash.
520 $a This dissertation focuses on feature extraction from natural images such that the extracted features are largely invariant under perceptually insignificant modifications to the image (i.e. robust). An iterative geometry preserving feature detection algorithm is developed based on an explicit modeling of the human visual system via end-stopped wavelets. For the second stage. I show that the decision version of the feature clustering problem is NP-complete. Then, for any perceptually significant feature extractor, I develop polynomial time clustering algorithms based on a greedy heuristic.
520 $a Existing algorithms for image/media hashing exclusively employ either cryptographic or signal processing methods. A pure signal processing approach achieves robustness to perceptually insignificant distortions but compromises security which is desirable in applications for multimedia, protection. Likewise pure cryptographic techniques while secure, completely ignore the requirement of being robust to incidental modifications of the media. The primary contribution of this dissertation is a joint signal processing and cryptography approach to building robust as well as secure image hashing algorithms. The ideas proposed in this dissertation can also be applied to other problems in multimedia security, e.g. watermarking and data hiding.
590 $a School code: 0227.
650 4 $a Engineering, Electronics and Electrical. $3 626636
690 $a 0544
710 2 0 $a The University of Texas at Austin. $3 718984
773 0 $t Dissertation Abstracts International $g 67-01B.
790 1 0 $a Evans, Brian L., $e advisor
790 $a 0227
791 $a Ph.D.
792 $a 2005
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3204165