我們把先把影像經過兩次拉式金字塔的拆解，然後把最高頻的部分作8 個方向的 2 維方向性濾波頻帶拆解，中間的頻率部分則是作4 個方向的 2 維方向性濾波頻帶拆解，
最低頻的部分不作方向性的拆解，整個拆解的架構如圖 19 左半部所示，而在頻域的分 解則如圖 19 右半部所表示。
圖 20 是把一張圖片用圖 19 的架構拆解後的結果，我們可以發現經過拆解後的次 頻帶，在相對應的方向上會有比較強的訊號出現。也就是說只要影像的邊緣(edge)和某 個頻帶一致性越大，在那個方向上就會有越強的訊號出現，顯示影像連續線條可以集 中在一個頻帶中，因此可以獲得較佳的壓縮效果。目前正在設計壓縮程序，希望近期 內會有具體成果。
圖 18 經過 4 個方向的 2 維方向性濾波頻帶所分解得到的頻譜
1
1 2
2 3
3 4
4 5
5 6
6 7
7
8
8 9
9
10
10 11
11
12 12
0 ω0
ω1
(π,π)
(-π,-π)
LP
LP DFB
4
DFB 8
圖 19 用來作輪廓轉換的架構和所對應的頻域的分解
16
參考文獻
[1] J. Bormans and K. Hill, “MPEG-21 Overview v.5,” ISO/IEC JTC 1/SC 29/WG 11/N5231, Shanghai, October 2002.
[2] Study of ISO/IEC 21000-4 FCD－IPMP Components, ISO/IEC JTC 1/SC 29/WG 11/N7426 July 2005, Poznan, Poland.
[3] Information Technology－Multimedia Framework (MPEG-21)-Part 5: Rights Expression Language, ISO/IEC 21000-5:2004, May 2004.
[4] C.A. Schultz, “Study of FPDAM ISO/IEC 14496-1:2001/ADM3,” ISO/IEC JTC 1/SC 29/WG11 N4849, Klagenfurt, July 2002.
[5] C.J. Tsai, M. van der Shaar and Y.K. Lim, “Working Draft 3.0 of ISO/IEC TR2100-12 Multimedia Test Bed for Resource Delivery,” ISO/IEC JTC1/SC29/WG11
MPEG2003/M10299, Hawaii, December 2003.
[6] Information Technology－Multimedia Framework (MPEG-21)-Part 12: Test Bed for MPEG-21 Resource Delivery, ISO/IEC 21000-12:2005, Apr. 2005.
[7] C.N. Wang, et al., “FGS-Based Video Streaming Test Bed for MPEG-21 Universal Mul-timedia Access with Digital Item Adaptation,” ISO/IEC JTC1/SC29/WG11
MPEG2003/M8887, October 2002.
[8] Information Technology－Multimedia Framework (MPEG-21)-Part 2: Digital Item Dec-laration, ISO/IEC 21000-2:2003, Mar. 2003.
1
1 2
2 3
3 4
4 5
5 6
6 7
7
8
8 9
9
10
10 11
11
12 12
0 ω0
ω1
(π,π)
(-π,-π)
0
1 2
4 3
5 6 7 8
12 11
10 9
圖 20 經過拉式金字塔和 2 維方向性濾波頻帶後拆解的圖片
17
[9] W3C (2001) XML Schema, http://www.w3.org/TR/2001//REC-xmlschema-1-20010502/
[10] C.-W. Fan, “MPEG-4 IPMPX Design and Implementation on MPEG-21 Test Bed,” M.S.
thesis, Dept. Electrical Engineering, National Chiao Tung University, Hsinchu, Taiwan, R.O.C., June 2004.
[11] X. Wang, et al., “An Example Implementation of MPEG-21 REL Reference Software,”
ISO/IEC JTC1/SC29/WG11 MPEG2003/M9581, March 2003.
[12] E. J. Cand`es. “Ridgelets: Theory and Applications”, Ph.D. Thesis, Department of Statis-tics, Stanford University, 1998.
[13] M. N. Do, Directional Multiresolution Image Representations, Ph.D. Thesis, Department of Communication Systems, Swiss Federal Institute of Technology Lausanne, November 2001.
[14] P. J. Burt and E. H. Adelson. “The Laplacian pyramid as a compact image code”, IEEE Transactions on Communications, vol. 31:4, pp 532–540, April 1983.
[15] S.-M. Phoong, C. W. Kim, P. P. Vaidyanathan, and R. Ansari. “A new class of two-channel biorthogonal filter banks and wavelet bases”, IEEE Transactions on Signal Processing, vol.
43:3, pp 649–665, Mar. 1995.
[16] A. Cohen, I. Daubechies, and J.-C. Feauveau. “Biorthogonal bases of compactly supported wavelets”. Commun. on Pure and Appl. Math., vol.45, pp 485–560, 1992.
[17] M. Vetterli. “Multidimensional subband coding: Some theory and algorithms”, Signal Processing, vol. 6:2, pp 97–112, Feb. 1984.
計畫成果自評
本計畫有以下幾類成果。第一類為MPEG-21 IPMP/REL System 與 Contourlet 所發展 出的技術、經驗及成品與國際 MPEG 標準直接相關，極具實用價值，可促進國內工業 研發技術開發。第二類為將上述技術提案至 MPEG 標準組織，有助我國技術之進入國 際舞台。第三類為計畫執行過程所獲得之研究成果論文四篇，已發表於國內外學術會 議。其四，參與計畫之同學可獲得國際多媒體最先進的MPEG-4 與 MPEG-21 相關技術 及多媒體系統設計經驗，畢業後進入產業，直接有助於產業界開發新產品，提昇我國 工業技術能力。達到人才培育之目的。
綜合評估：本計畫產出相當多具有學術與應用價值的成果，特別是直接參與國際標 準會議，在國際上展示成果。並培育高科技人才培育，整體成效可稱良好。已發表(含 接受)學術論文五篇，博士學位論文一冊，以及碩士學位論文二冊如下表。
18
Publications
(1) F.-C. Chang and H.-M. Hang, “A relevance feedback image retrieval scheme using multi-instance and pseudo image concepts,” IEICE Transsactions on Information and
Systems, pp.1720-1731, May 2006. (SCI, EI) [NSC 91-2219-E-009-041]
(2) F.-C. Chang, H.-C. Huang and H.-M. Hang, “Layered access control schemes on water-marked scalable media,” J. VLSI Signal Proc Systems for Signal, Image, and Video
Technology, to be published. (SCI, EI) [NSC 92-2219-E-009-008]
(3) C.-H. Lu, F.-C. Chang, and H.-M. Hang, “Design and Implementation of MPEG-21 IPMP and REL on MPEG-21 Testbed using IPMPX Framework”, in 2006 Conf. on
Computer Vision, Graphics, and Image Processing, Tao-yuan, Taiwan, Aug. 2006.
(4) C.-H. Lu, F.-C. Chang, and H.-M. Hang, “A Content Protection Scheme Using MPEG-21 Concepts and Tool,” in Europe-China Conference on Intellectual Property in Digital
Media (IPDM), Shanghai, Oct. 2006.
(5) Y.-T. Shih, F.-C. Chang, and H.-M. Hang, “A Digital Content Protection Scheme Using MPEG-21 REL with Applications to DVB systems,” in 2006 IEEE International
Con-ference on Intelligent Information Hiding and Multimedia Signal Processing
(IIHMSP-2006), Pasadena, Dec. 2006.
(6) Feng-Cheng Chang 張峰誠, Digital Image Retrieval and Scalable Media Protection, Ph.D. Dissertation, NCTU, May 2006
(7) Ying-Tzu Shih 施瑛姿, A Digital Content Protection Scheme Using MPEG-21 REL with
Applications to DVB Systems, MS Thesis, NCTU, November 2005.
(8) Chia-Hsien Lu 呂家賢, An Implementation of MPEG-21 IPMP and REL with MPEG-4
IPMPX Framework on MPEG-21 Testbed, MS Thesis, NCTU, June 2006.
DESIGN AND IMPLEMENTATION OF MPEG-21 IPMP AND REL ON MPEG-21 TESTBED USING IPMPX FRAMEWORK
Chia-Hsien Lu (
呂家賢), Feng-Cheng Chang (
張峰誠), and Hsueh-Ming Hang (
杭學鳴) Dept. of Electronics Engineering, National Chiao Tung University
[email protected], [email protected], [email protected]
ABSTRACT
Due to the fast advances in consumer electronics and broadband networks, people nowadays can easily create and distribute digital contents. How to effectively manage and protect the rights of consuming digital contents becomes an important issue. In this paper, the technologies of MPEG-21 IPMP (Intellectual Property Management and Protection) and REL (Rights Expression Language) are adopted to construct a DRM system. The MPEG-4 IPMPX (Intellectual Property Management and Protection Extension) system is also used as the basic framework in our implementation. The MPEG-21 IPMP provides ways to protect a digital element. The MPEG-21 REL is able to describe various kinds of rights and it provides an authorization model to generate an authorization proof to manage the rights. To implement these standard specifications, we design a set of IPMP Tools, which carry the functionalities of the MPEG-21 IPMP and REL. We then integrate this set of tools into the MPEG-4 IPMPX framework on the MPEG-21 Test Bed. At the end, we develop three application examples to demonstrate that our system can successfully safeguard digital resources and effectively manage the rights.
1. INTRODUCTION
As the network and digital media technologies advance, it is easy nowadays for everyone to create and distribute digital multimedia contents. The intellectual property protection and management becomes an important issue.
Therefore, modern multimedia system designs often include digital rights management (DRM) as an essential component.
To protect digital contents from eavesdropping, encryption is an important tool. Thus, most of the popular DRM systems incorporate the encryption functionality. However, encryption alone is not sufficient to provide a more sophisticated service such as a right that authorizes to the users to consume the resource under specific conditions or constraints. A complete
DRM system requires a means to describe who owns a piece of content and how it can be used. This can be achieved by incorporating a rights expression language.
Generally speaking, a rights expression language is able to describe the rights holder of a certain resource, the target consumer of the resource, the allowed rights on the resource, and the necessary conditions for resource consumption. There are a few rights expression languages defined for various application fields. One of them is defined by the MPEG committee for the use in connection with the MPEG-21 multimedia framework.
MPEG-21 [1] is a set of specifications defined as
“a multimedia framework to enable transparent and augmented use of multimedia resources across a wide range of networks and devices used by different communities.[1]” There are two parts in MPEG-21 that constitute the basis of a DRM system. The MPEG-21 Part 4 [2] defines the high level concepts and schema for Intellectual Property Management and Protection (IPMP); and the MPEG-21 Part 5 [3] defines the Rights Expression Language (REL).
The MPEG-21 REL is an XML-based language that can declare an authorized distribution for the use of any content, resource, or service owned by specific users.
It provides flexible, exact, and rich representation of rights. It can be used in various applications due to the interoperability of this language. According to a specific application, users who use REL for rights management can define their own extensions as well as create a specific profile. The other parts of MPEG-21 are also related to specific aspects of a DRM system.
In this paper, we study and implement the MPEG-21 IPMP and REL specifications. By analyzing several typical multimedia DRM scenarios, we design a DRM scheme to protect the digital assets and to perform the rights management. Its functionality is compliant to the MPEG-21 IPMP and REL concepts. Because the specifications in MPEG-21 IPMP are abstract, we map its DRM functionality to the MPEG-4 IPMP Extension (IPMPX) [4] architecture. To verify the feasibility of our design, we choose the MPEG-21 Test Bed [5][6][7] as the content delivery platform. In addition to the ability of simulating a basic real-time content delivery system,
760
2006 19th IPPR Conference on Computer Vision, Graphics and Image Processing
it also has a simple implementation of the IPMPX architecture inside. Hence, we implement our DRM by integrating the MPEG-21 IPMP, the MPEG-21 REL, together with IPMPX on the Test Bed.
This paper is organized as follows. We first introduce the concepts and specifications of the MPEG-21 IPMP in Sec. 2. Second, the REL concepts and specifications are discussed in Sec. 3. Then, Sec. 4 gives a brief overview of the MPEG-21 Test Bed with IPMPX.
In Sec. 5, we describe the details of our design and implementation of the MEPG-21 IPMP and REL within the MPEG-4 IPMPX framework on the MPEG-21 Test Bed. Finally, we design three application examples to demonstrate the functionalities of our designed DRM system. A few conclusion remarks of our work are given in Sec. 7.
2. MPEG-21 IPMP
The goal of the MPEG-21 IPMP is to provide the management of rights and intellectual property through the use of protected Digital Items. A Digital Item (DI) is the subject that the MPEG-21 technology applies to. It can be a multimedia clip, a document, or even a service which provides sort of “digital content”. To describe the properties of a DI, the MPEG-21 developed the Digital Item Declaration Language (DIDL), which is an XML-based expression defined in the MPEG-21 Digital Item Description (DID) specifications [8].
A DID is a plain-text description format of a DI.
However, in many multimedia protection schemes, not only the content itself but also the meta-data of the content should be protected against unauthorized access.
Therefore, an additional secure representation of DI is developed. A protected representation of DID Model Structure is proposed as IPMP DIDL.
Both representations (in DIDL and in IPMP DIDL) are semantically equivalent. They refer to the identical digital item by different syntax. For every DIDL element, there is an interchangeable IPMP DIDL element which carries valuable content (as shown in Fig. 1). The carried content can be encapsulated in either the unprotected or the protected form. For a protected element, the additional side information is required in order to recover it to its original (unprotected) form. In the specifications, the IPMP Information Descriptor and IPMP General Information Descriptor are defined for representing the necessary side information including tools, mechanisms, and licenses.
Since the IPMP DIDL shares a vast amount of definitions with the DIDL, the hierarchical relationships are illustrated in Fig. 2. For an IPMP capable consumer, it reads in the description using the IPMP DIDL schema.
This schema consists of the definitions derived from the generic DID model, the definitions from the DIDL schema, and its own IPMP specific definitions.
3. MPEG-21 REL
The MPEG-21 REL is an XML-based and machine-interpretable rights expression language that declares an authorized distribution for the use of any content, resource, or service owned by specific users. It defines an unambiguous syntax to specify rights.
To fulfill the demand of distribution and protection of digital contents, it is essential that the language provides an authorization model for checking whether the rights specified in the license is validate. To be able express a wide variety of business models and to enable multimedia distribution and usage of all types of digital resources, the MPEG-21 REL is designed to be extensible and flexible.
Fig. 3 Data model of an REL license.
<Container>
Fig. 1 Element interchangeability.
Fig. 2 Relationship between DID and IPMP DID.
3.1. Data Model
The most important data model in the MPEG-21 REL is the license structure. As shown in Fig. 3, a license may contain any number of issuers and grants. An issuer is a subject who issues the license. It could be a person, an organization, or a license service. A grant represents a permitted combination of the digital rights management entities.
A grant is constructed by four components. The only mandatory component is the Right, which specifies the allowed action. The Principal component specifies the target to which the license is issued. The Resource component specifies the target which the action is applied to. The Condition component specifies the criterion for deciding the validity of the grant.
In a syntactical expression, the Principal is the subject; the Right is the verb; the Resource is the object; and the Condition is the terms, conditions, and obligations under which the right can be exercised.
Hence, a grant is equivalent to the sentence: “under the Condition, the Principal is allowed to exercise the Right on the Resource.”
3.2. Authorization Model
To determine the permission of exercising a certain right, an authorization model is defined in the MPEG-21 REL.
As shown in Fig. 4, an authorization proof is the binary result (true or false) of testing the authorization request against the authorization story.
An authorization request is composed of several items. They form a structure to express the question: “is it permitted for the principal to perform the right upon the resource during the time interval, according to the authorization context, the set of obtained licenses, and the set of trusted grants?” The authorization context is used to hold additional properties of the request, to be matched with the conditions of the grants.
Upon receiving the request, the REL tool extracts the licenses and grants to construct an authorization story. Each license is used to construct an authorizer (which may contain a story), and the extracted grants are put into the pool of authorized grants. Then, additional authorized grants are derived from the authorizer, recursively. The primitive grants are derived from the grants, the result is true. Otherwise, the proof is failed.
3.3. Extensibility and Profiling
Because the syntax of the REL are described using the XML Schema [9], the MPEG-21 REL can offer a high
degree of flexibility in its extensibility. The schema can be organized into three categories: core, standard extension, and multimedia extension (Fig. 5).
The core schema defines the general concepts that form the basic architecture of the language, particularly the ones strongly related to a trust evaluation. The standard extension schema defines the concepts that are generally and broadly useful and applicable to DRM scenarios. The multimedia extension schema defines the DRM concepts specifically related to the multimedia content such as books, video, and audio. Both the standard and the multimedia extensions are extended from the core schema. The REL allows, but not limited, to express rights using the defined schemas. Future applications may extend the core schema to define new expression elements.
The extensibility enables the REL to cooperate with a wide range of applications. However, not all applications require the full support of REL. Especially for an embedded application, the limited resource of the device would restrict the DRM functionality to a subset of the REL elements. The profiling process of the REL is to select a set of items for a specific purpose. This process produces a subset of the language, and the REL peers are able to determine whether they are interoperable or not.
Fig. 4 MPEG-21 REL authorization model.
Core
Fig. 5 REL Extensibility.
762
4. MPEG-21 TESTBED
The MPEG-21 Test Bed is chosen as our implementation platform. We briefly describe the structure of the Test Bed and the existing MPEG-4 IPMPX on it in this section.
4.1. MPEG-4 IPMPX
The MPEG-4 IPMPX system became an ISO standard in October 2002. It provides the functionality to protect and manage the MPEG-4 contents. The MPEG-4 others. A message has three basic fields. The type field denotes the type of the message; the source field denotes the component which sends the message; and the destination field denotes the recipient of the message.
To send a message, the originator passes the message to the Message Router. Then, the MR transports the message to the destination component.
The interaction to and from the MR are through the message interface. The Tool Manager manages the lifecycle of an IPMP Tool. When a tool is accessed, the TM is responsible for the instantiation, initialization, and connection to the specified control point. After the use of a Tool, the TM disconnects, uninitializes, and destructs it.
In an MPEG-4 elementary stream data path, there exist several control points. A control point is a virtual container in which the triggered IPMP Tools process the content one after another. For the tools that do not associate with a control point, they act like a service in the system. We will describe a design of REL service in Sec. 5.
4.2. MPEG-21 Test Bed with IPMPX
The purpose of the MPEG-21 Test Bed is to provide a flexible and fair test environment for evaluating delivery technologies for MPEG contents over IP network. The
The purpose of the MPEG-21 Test Bed is to provide a flexible and fair test environment for evaluating delivery technologies for MPEG contents over IP network. The