a web-based melanoma image diagnosis support … topic occurrence speciﬁes information, ......

Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

A web-based melanoma image diagnosis support systemusing topic map and AJAX technologies

A. PAPASTERGIOU1, P. TZEKIS2, A. HATZIGAIDAS1, G. TRYFON1,

D. IOANNIDIS1, Z. ZAHARIS1,3, D. KAMPITAKI1 & P. LAZARIDIS1,4,

1Department of Electronics, Alexander Technological Educational Institute of Thessaloniki, 57400

Thessaloniki, Greece, 2Department of Sciences – Faculty of Mathematics, Alexander Technological

Educational Institute of Thessaloniki, 57400 Thessaloniki, Greece, 3Telecommunications Center, Aristotle

University of Thessaloniki, 54124 Thessaloniki, Greece and 4Departement Communications et

Electronique, Unite de Recherche Associee au Centre National de la Recherche Scientifique, 820 Ecole

Nationale Superieure des Telecommunications, 46 rue Barrault, 75634 Paris Cedex 13, France

AbstractThe design and implementation of a web-based diagnostic support tool for melanoma dermatologicalimages and related diagnostic data is presented. The proposed system is semantic web-based and isdriven by exploiting the combination of AJAX framework and topic map technology. A novel client/server architecture was developed that enables several clients to interact online with the topic map-basedsystem. Users have the ability to access the system anywhere and anytime via a simple Internet browser.Additionally, an ABCD application has been developed for automated calculation of ABCD parametersand consequently embedded in the proposed TM-based system.

Keywords: Topic maps, AJAX, semantic web, melanoma, dermoscopy, ABCD rule

1. Introduction

Melanoma is one of the most aggressive types of skin cancer [1–3]. Early diagnosis is the most

reliable solution for an effective treatment of melanoma [4–6]. There is a continuous research

worldwide to support dermatologists with effective diagnostic tools and methods. Dermo-

scopy (dermatoscopy, epiluminescence microscopy) has been established as a non-invasive

method for improving diagnosis of melanoma [7–11]. The dermoscopic diagnosis of

melanoma is based on various analytic approaches and algorithms that have been set forth in

the last few years [12–14]. The ABCD rule (A (asymmetry), B (border), C (color), D

(diameter or differential structures)) is a standard used in dermatoscopy analysis for

classification of dermatological images to benign, suspicious or melanoma [13,15].

There has been a great deal of scientific research that aims at providing reliable diagnostic

support to dermatologists through computerized means, towards early diagnosis of

Correspondence: Panayotis Tzekis, Department of Sciences – Faculty of Mathematics, Alexander Technological Educational

Institute of Thessaloniki, P.O. Box 141, GR – 574 00, Thessaloniki, Greece. Tel: þ30-2310-791611. Fax: þ30-2310-791622.

E-mail: [email protected]

Informatics for Health & Social Care

June 2008; 33(2): 99 – 112

ISSN 1753-8157 print/ISSN 1753-8165 online � 2008 Informa UK Ltd.

DOI: 10.1080/17538150802127256

Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

melanoma. Computerized methods increase diagnostic accuracy for dermatologists and

enable storing of images with diagnostic information for further investigation or creation of

new diagnosis methods [16–23]. Furthermore, the increasing amount of digital dermoscopic

images produced in medical institutions helps to develop new advanced image retrieval

systems for more efficient information storage and management, and also for diagnostic and

teaching support [24,25].

One of the major requirements of such a system is the ability to assure interoperability and

reusability of images and related diagnostic data. Research is turning towards producing not

stand-alone, but web-based applications that make information more accessible and

interoperable by all the users involved. The recent network technology of mobile and satellite

communications has significantly improved the accessibility of information resources.

Additionally, such a system should be capable of demonstrating some form of knowledge-

based reasoning and decision support [26–29]. Semantic web technologies, such as ontologies

and topic maps, seem to have a great deal of potential to provide explicit representation of the

semantics of data in order to achieve interoperability and reusability of data as well as

advanced reasoning support in such a system [30–32].

Taking into consideration the above, the main target of this work is to provide an integrated

semantic web-based system that manages and retrieves information from digital melanoma

images and associative data. The objective of this paper is to present in detail the design and

implementation of the proposed system, which is structured by exploiting the combination of

AJAX framework and topic map technology. The ultimate goal of our work is to investigate

the potential benefits of encoding medical information using topic map technology, in order

to create an effective diagnostic support tool for melanoma cases.

2. Using topic map technology

As mentioned before, there has been a steady increase in malignant melanoma incidents

worldwide. In order to support early diagnosis, new tools and methods are needed that

manipulate the increasing amount of visual information produced in that field. The scope of

this work is to exploit the potential advantages and difficulties of using the topic map standard

for managing melanoma case images and their related data, in order to create a web-based

medical application that provides diagnosis support utilities to dermatologists.

Topic map (TM) is an ontology technology, designed for information and knowledge

organization [33–35]. By providing a standard interchange syntax specification [36], TM

intends to support exchange, reusability and interoperability of knowledge among different

applications. TMs are constructed around topics that represent subjects. Associations are

used to describe n-ary relationships between topics and they are generally categorized

according to association types. Topics that participate in an association, play roles referred to

as ‘association roles’. A topic occurrence specifies information, which is related to a given

subject. If the subject is a resource that can be addressed, the resource address can be used as

a subject locator [33–39].

TMs have been designed to be able to manage and present any kind of subjects and any

relationship between the subjects in any kind of ontological context [40,41]. A TM can build

a domain knowledge model above related resources data, thus achieving the clear separation

between knowledge and information layers. In this way, the TM enables advanced navigation

and information retrieval in complex datasets [37–39]. Although a relational database is a

reliable solution, it has to be rewritten in order to support a new class of entity and queries. By

using TMs, the user needs only to alter properties and relationships, without having to rewrite

program code or queries [37].

100 A. Papastergiou et al.

Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

Finally, a functional knowledge repository should allow effective search and retrieval of

information. Ontopia has developed Tolog 1.0 for querying TMs [42]. Tolog is a Prolog-like

language based on first-order predicate logic. Complex queries can be expressed by applying

Boolean variables and built-in predicates supporting in this way querying on both data and

ontology [43]. Furthermore, the users can define inference rules in a separate file that gets

included on the execution of a query. Then, inference rules can be used to simplify queries

and to provide a way to capture implicit relationships through the declaration of simple rules

[42–44].

3. Incorporating AJAX framework

A TM-based information management and retrieval system suitable for possible melanoma

images has been developed during our research [45]. The programming language used is Java

and in particular the Netbeans IDE platform [46]. The initial evaluation of the system shows

promising results. Nevertheless, some problems have been identified during the evaluation

and they are reported below.

The communication between client and server is quite slow. The uploading of images is

rather time-consuming. In order to interact with the system and grant access to the server, the

users have to install Ontomel application into their PCs. This procedure seems to be rather

fuzzy for users, especially if they are not familiar with computers. A viable solution could be

access to the system by using a common Internet browser like Internet Explorer. Netbeans

environment provides many tools to programmers for creating Windows-based applications.

Nevertheless, in practice the application produced was rather complicated. Thus, an attempt

was made to simplify the programming process in order to create easy-to-use client interfaces.

The above problems can be solved and further improvement may be achieved by employing

AJAX technology. The proposed system is web-based and is structured by exploiting the

combination of AJAX framework and TM technology. A novel client/server architecture was

developed that enables several clients (e.g. dermatologists) to interact online with the TM-

based system. This framework allows the end users to view and modify huge amounts of

medical information as well as to share the data. Web applications have many benefits over

desktop applications; a larger audience can be reached, they have easier installation and

support and also easier development. However, Internet applications are not always as

appealing and user-friendly as traditional desktop applications. With AJAX, Internet

applications can be made more attractive and user-friendly. AJAX stands for Asynchronous

JavaScript and XML and is a web development technique for creating interactive web

applications. AJAX is a promising solution for closing the gap between web applications and

desktop applications regarding usability and responsiveness [47–50]. The AJAX web

application model enhances the interaction of the user with the user interface by introducing

an intermediary (AJAX engine) between the user and the server. In brief, instead of loading a

web page at the start of each session, the browser loads an AJAX engine that is responsible for

both rendering the user interface and communicating asynchronously with the backend servers

[49].

4. Architecture of the proposed web-based topic map system

This section presents the framework developed for the generation of the web-based diagnostic

support tool. The core of the proposed system is the Enterprise Topic Map Server (ETMS),

which has been implemented using several Java servlets (user maintenance, ontology,

information storage and retrieval, etc.). ETMS administrates the Ontomel TM and

Web-based melanoma image diagnosis support system 101

Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

coordinates communication issues with client application and MySQL database [45,51].

Communication issues and exchange of TM are established using the proposed AJAX

framework via simple HTTP protocol requests, as illustrated in Figure 1.

Administration of TMs is made by ETMS using TM4J library. TM4J provides an

Application Programming Interface (API) to allow programmers to create and modify TM

structures [52]. The TM4J engine provides interfaces for querying TM structures using the

Tolog query language and parsing TMs from XTM syntax files or writing TMs to XTM

syntax files [42,52]. Finally, TM4J library manages TMs that are persistently stored in the

relational database MySQL using the Hibernate O-R mapping [52].

4.1. Description of the overall system

The Ontomel framework is provided to the end-users who just want to access the system

using a simple web interface but they do not have an implicit knowledge of TM technology.

Specifically, simple form-based graphical user interfaces (GUIs) have been implemented

Figure 1. Client/server architecture using AJAX framework.


Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

using the Google web toolkit (based on AJAX framework) (Figure 2) [53]. Additionally, a

software tool, called Yellow-TM Editor, is addressed to administrators of the system by the

server. Yellow-TM tool was implemented during this research, aiming to provide an

advanced tool for TM authoring [54]. The system administrator manages information using

Yellow-TM and thus communicates with the TM-based server. Yellow-TM Editor enables

authoring and visualization of TM, as well as definition of TM Schema and expression of

rules and queries in Tolog [42,54].

4.2. Creating TM-based ontology

The present subsection describes in detail the creation of a functional TM-based knowledge

base that contains digital dermoscopic images of possible melanoma cases, as well as data

related to images, doctor reports, diagnostic findings and patient information (Figures 1

and 2).

The first step is to identify the basic components that have to be included in order to create

a knowledge infrastructure of digital images and clinical data. Figure 3 depicts the basic

structure of the proposed database. Each doctor will be associated to a patient. Then the

patient is associated to a ‘concept’ melanoma and each melanoma is associated to a digital

image. This is the main backbone of the database. Appropriate attributes must be defined for

all components, so that the components can be annotated efficiently and describe the

knowledge that they inherit. The properties of images are described by a set of variables

required for the calculation of the TDS (Total Dermatoscopy Score) index [15], which is an

efficient tool widely used in melanoma diagnosis. The calculation of TDS is based on four

features, which are the ‘Asymmetry’, the ‘Border’, the ‘Colors’ and the ‘Diameter’. The value

of each feature is multiplied by a given weight factor to yield the value of TDS according to

the following formula:

TDS ¼ 1:3 � Aþ 0:1 � Bþ 0:5 �Cþ 0:5 �D ð1Þ

TDS contributes to the differentiation between benign and malignant lesions. Values of TDS

less than 4.75 correspond to a benign skin lesion, values of TDS between 4.75 and 5.45

Figure 2. Representation of TM-based system using a web user interface and AJAX framework to communicate with

the TM-based server.


Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

correspond to a suspicious skin lesion, and finally values greater than 5.45 indicate that the

skin lesion is a melanoma with a high degree of certainty. The attributes attached to every

image are the date, the actual resource path of the photo, the ABCD values, the value of TDS,

a possible diagnosis, and the results from the laboratory exams. Accordingly, for every patient,

properties like name, age and gender can be defined.

A great advantage of TMs is the clear separation between the description of the information

structure (ontology and knowledge layer) and the physical information resources (information

layer) (Figure 4). In a TM, the ontology layer contains topic types, association types, role

types and occurrence types while the knowledge base layer contains instances of the types

defined in the ontology layer [18]. In Figure 4, the ontology layer outlines the types that must

be established in order to encode the basic structure of the database presented in Figure 3

using the TM standard.

The Yellow-TM editor has been used to construct the knowledge repository based on the

TM model that contains melanoma images and related data. The Yellow-TM editor is a

form-based editing environment for the administration of the system [54,55]. The Yellow-

TM tool displays visually the TM concepts providing a dynamic graphic representation

(knowledge layer in Figure 4). Topics and associations of the TM are represented as nodes,

Figure 3. Basic structure of the proposed topic map database.


Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

while roles are represented as arcs that connect topics and associations. By right clicking on

the desired topic an editable ‘Properties’ panel appears showing all the corresponding

properties. In detail, for a topic of topic type ‘image’ the actual resource address can be used

as a subject while the date, the ABCD values and the TDS value are defined as occurrences. It

must be noticed, that this is a preliminary database and only a basic design is attempted.

However, the database can be easily expanded and enhanced with many more components

and properties. TMs enable scalability and expandability of the implemented database, while

Figure 4. Clear separation among ontology, knowledge and information layers.


Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

there is no need to alter the basic schema of the database. Thus, the program enables the user

to add a new type of information, while a ‘topic type’ that corresponds to this information is

created and employed in the TM. In this way, the system can be suitably modified depending

on the user needs. Additionally, the Yellow-TM tool supports the input of queries and rules

based on Tolog language [54]. Suitable rules and queries have been established using Tolog

language [42] in order to enable the user to retrieve knowledge contained within the

implemented TM and the associated dermoscopic images.

5. ABCD application

The main demand that seems to be indispensable for a diagnostic support tool of melanoma is

the automatic detection of ABCD values. Therefore, an ‘ABCD’ program for automated

calculation of ABCD values has been designed and incorporated into the TM-based system.

The ABCD application is executed by the ETMS (Figure 5). When a user uploads an image

to the system, the ABCD parameters are automatically estimated and attached to the specific

dermatological image.

According to similar programs [23,25], the ABCD program (Figure 5) must perform the

following steps:

. segmentation of image in order to separate the tumor area from the surrounding healthy

skin,

. feature extraction for calculation of ABCD values and TDS,

. return of the estimated diagnostic values to the user.

The following subsections outline the image segmentation and feature extraction issues.

Figure 5. Schematic description of ABCD application in the TM-based system.


Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

5.1. Image segmentation issues

According to the design analysis, the first step towards the implementation of an ABCD

program is the definition of tumor area covered by melanoma and its extraction from the

healthy skin. A new and fast algorithm has been developed in our laboratory (submitted for

publication). This algorithm defines the lesion boundary in order to distinguish accurately the

area of the melanoma from the rest of the skin in reasonable computing time.

5.2. Feature calculation

Assuming that the area of possible melanoma is estimated, several meaningful features can be

extracted [21]. The ABCD calculation program is based on the ABCD dermoscopic

classification rule. The features to be extracted from the images are the A, B, C and D values

that correspond to the ABCD mnemonic rule [15].

In the proposed approach we chose four descriptors to represent the clinical features

included in the ABCD rule, as follows: (1) derivation factor for Asymmetry; (2) perimeter

factor for Border; (3) color presence for Color; and (4) the greatest diameter (in mm) of the

lesion for Diameter.

5.2.1. A: Derivation factor. The derivation factor is estimated by the expression:

s2 ¼

Pni¼0

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffixi �mxð Þ2þ yi �my

� �2q

� r

� �2

n� 1ð Þr

where n is the number of points of the lesion border, r is the radius of the circle that has the

same area as that of the lesion, (mx,my) is the center of gravity of the border and the circle and

(xi,yi) is the ith point of the border. If the derivation factor is close to zero, the melanoma is

almost perfectly circular.

5.2.2. B: Perimeter factor. Regarding border irregularities, the perimeter factor is calculated by:

Pf ¼n

2pr

where n is the number of points of the perimeter. When the Pf is close to 1, the perimeter is

‘smooth’.

5.2.3. C: Color presence. The presence of six main colors (white, red, light brown, dark brown,

blue-gray, black), is considered. The presence of these six basic colors inside the lesion gives

one point to each color for the calculation of the C score (0 – 6 points) [13,15]. A 66 6 item

table is defined. Each row of the table contains the RGB lower and upper limit values that

correspond to each one of the six colors. The C score is increased by one point if there is at

least one pixel inside the test region whose RGB values are between the RGB limit values of

any of the six colors.

5.2.4. D: Greatest diameter. The greatest diameter of the region is simply estimated in pixel

coordinates by calculating the largest distance between the contour points of the region.


Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

The ABCD application is implemented in C language and has been incorporated in the

TM-based system. The program calculates the TDS according to the ABCD estimated

values. Consequently, ETMS encodes these values and input data related to the image as TM

structures and saves them in the system database.

6. Describing the Ontomel application

Registered users log into the system using an Internet browser anywhere and anytime, and

subsequently interact with it (Figure 2). The user may upload a dermatological image, related

data, patient data and clinical data in a simple way (Figure 6). By using the AJAX engine,

several calls (asynchronous) can be established with the ETMS that resides in a Tomcat

server. The ETMS encodes all input data according to the corresponding TM structure and

saves them in the TM-based database.

When a user uploads a new dermoscopic image from his computer to the ETMS, the

ABDC application automatically estimates the ABCD values. After the segmentation process

and the ABCD calculation, the ABCD values are presented on the right of the image in a

simple numerical output. In the background, ETMS encodes the estimated ABCD values,

the uploaded image and all related input data as TM structures, which are saved in the

MySQL database.

Moreover, enhanced navigational and retrieval functionalities are provided by easy-to-use

GUIs. The user can locate a patient from a predefined list and see associated melanoma

Figure 6. The user uploads a new image and the system returns ABCD values.


Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

images and clinical information (Figure 6). Additionally, the end-user can use the predefined

rules and queries established in the system, for a more sophisticated access and retrieval of

information, e.g. to find melanoma images with similar ABCD features. However, from the

end-user perspective, the client application has to minimize the required user input. Also, the

user probably will not have knowledge of TM or Tolog syntax. This problem is solved

by providing an effective and easy-to-use GUI that runs the predefined rules and queries

(Figure 7).

The URL of the implemented system is http://195.251.241.43:8080/WebOntomel/

webontomel.Main/index.html. Researchers that are interested are encouraged to visit it and

explore all its available features. Potential user opinions and evaluations will provide valuable

feedback on how the overall system should be refined and integrated.

7. Conclusions

The design and implementation of a web-based diagnostic support tool for melanoma

dermatological images and related diagnostic data was presented. The user can access the

system via a simple Internet browser, thus exploiting the benefits provided by the modern

network technology of mobile and satellite communications. The proposed system

incorporates AJAX framework and TM technology. TM technology seems to offer the

necessary framework for knowledge management to support share ability and reusability of

information as well as advanced reasoning support. Also, AJAX technology offers the

possibility of creating web applications that resemble desktop applications in terms of usability

and responsiveness. Additionally, a calculation program for automatic estimation of the

ABCD parameters of uploaded images has been implemented and incorporated into the

proposed web-based system.

Figure 7. The user retrieves information.


Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

The main scope of this work is to explore the potential benefits of encoding information

using TM technology and consequently to provide an efficient and integrated tool as

diagnostic aid to dermatologists treating melanoma cases. Currently, the feasibility and

efficiency of the proposed system is under evaluation and guidelines for future work are

outlined.

Declaration of interest: The authors report no conflicts of interest. The authors alone are

responsible for the content and writing of the paper.

References

1. Diepgen TL, Mahler V. The epidemiology of skin cancer. Br J Dermatol 2002;146:1–6. Available: http://

www.blackwell-synergy.com/doi/pdf/10.1046/j.1365-2133.146.s61.2.x via the INTERNET.

2. Koh H. Cutaneous melanoma. N Engl J Med 1991;325:171–182.

3. The Skin Cancer Foundation [cited 2004 Oct 12]; Available from: http://www.skincancer.org/self_exam/

look_for.html.

4. Friedman RJ, Rigel DS, Kopf AW. Early detection of malignant melanoma: the role of physician examination and

self-examination of the skin. CA Cancer J Clin 1985;35:130–151.

5. Friedman RJ, Rigel DS, Silverman M, Kopf AW. The continued importance of early detection of malignant

melanoma. CA Cancer J Clin 1991;41:201–226.

6. Menzies SW. Cutaneous melanoma: making a clinical diagnosis, present and future. Dermatol Ther 2006;19:32–

39. Available: http://www.blackwell-synergy.com/doi/pdf/10.1111/j.1529-8019.2005.00054.x via INTERNET.

7. Argenziano G, Soyer HP. Dermoscopy of pigmented skin lesions—a valuable tool for early diagnosis of

melanoma. Lancet Oncol 2001;2:443–449.

8. Binder M, Schwarz M, Winkler A, et al. Epiluminescence microscopy: a useful tool for the diagnosis of

pigmented skin lesions for formally trained dermatologists. Arch Dermatol 1995;131:286–291.

9. Johr R, Menzies S. Lesions on dermoscopy. Dermatol Surg 2001;27:911–912.

10. Menzies SW. A method for the diagnosis of primary cutaneous melanoma using surface microscopy. Dermatol

Clin 2001;19:299–305.

11. Zsolt BA. Dermoscopy (epiluminescence microscopy) of pigmented skin lesions. Dermatol Clin 1997;15:79–95.

12. Consensus Net Meeting on Dermoscopy. Available from: http://www.dermoscopy.org/default.asp.

13. Johr HR. Dermoscopy: alternative melanocytic algorithms—the ABCD rule of dermatoscopy, Menzies scoring

method, and 7-point checklist. Clin Dermatol 2002;20:240–247.

14. Consensus Net Meeting on Dermoscopy (CNMD). Unifying concepts of dermoscopy. Available from:

www.dermoscopy.org.

15. Stolz W, Braun-Falco O, Bilek Landthanler M, Conneta AB, Riemann A. ABCD rule of dermatoscopy: a new

practical method for early recognition of malignant melanoma. Eur J Dermatol 1994;4:521–527.

16. Blum HL, Ellwanger U, Schwabe R, Rassner G, Garbe C. Digital image analysis for diagnosis of cutaneous

melanoma. Development of a highly effective computer algorithm based on analysis of 837 melanocytic lesions.

Br J Dermatol 2004;151:1029–1038. Available: http://www.blackwell-synergy.com/doi/pdf/10.1111/j.1365-

2133.2004.06210.x via the INTERNET.

17. Hall PN, Claridge E, Morris Smith JD. Computer screening for early detection of melanoma—is there a future?

Br J Dermatol 1995;132:325–338.

18. Horsch A, Stolz W, Neiss A, et al. Improving early recognition of malignant melanomas by digital image analysis

in dermatoscopy. Stud Health Technol Inform 1997;43:531–535.

19. Gewirtzman AJ, Braun RP. Computerized digital dermoscopy. J Cosmet Dermatol 2003;2:14–20. Available:

http://www.blackwell-synergy.com/doi/pdf/10.1111/j.1473-2130.2003.00009.x via the INTERNET.

20. Day GR, Barbour RH. Automated melanoma diagnosis: where are we at? Skin Research and Technology

2000;6:1–5. Available: http://www.blackwell-synergy.com/doi/pdf/10.1046/j.1365-2133.2002.04978.x via the

INTERNET.

21. Maglogiannis I, Kosmopoulos D. Computational vision systems for the detection of malignant melanoma.

Oncology Reports special issue on Computational Analysis and Decision Support Systems in Oncology

2006;15:1027–1033.

22. Sboner A, Eccher C, Blanzieri E, Bauer P, Cristofolini M, Zumiani G, Forti S. A multiple classifier system for

early melanoma diagnosis. Artificial Intelligence in Medicine 2003;27:29–44.


Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

23. Schmid-Saugeon P, Guillod J, Thiran JP. Towards a computer-aided diagnosis system for pigmented skin

lesions. Computerized Medical Imaging and Graphics 2003;27:65–78.

24. Kamel Boulos MN. A preparatory study on the Dermatology Virtual Branch Library of the NeLH [MSc Thesis

in Medical Informatic]. London: GKT School of Medicine, KCL, University of London; 2000.

25. Oka H, Hashimoto M, Iyatomi H, Tanaka M. Internet-based program for automatic discrimination of

dermoscopic images between melanoma and Clark naevi. Br J Dermatol 2004;150:1041.

26. Tang LHY, Hanka R, Ip HHS. A review of intelligent content based indexing and browsing of medical images.

Health Informatics Journal 1999;5:40–49.

27. Muller, H, Michoux, N, Bandon D, Geissbuhler A. A review of content-based image retrieval systems in medical

applications—clinical benefits and future directions. International Journal of Medical Informatics 2004;73:1–30.

28. Iyatomi H, Oka, H, Hashimoto M, Tanaka M, Ogawa K. An internet-based melanoma diagnostic system—

toward the practical application. Computational Intelligence in Bioinformatics and Computational Biology

(CIBCB); 2005. Proceedings of the 2005 IEEE Symposium;2005. p. 1–4. Available: http://ieeexplore.ieee.org/

iel5/10629/33563/01594952.pdf?tp=&isnumber=&arnumber=1594952 via the INTERNET.

29. Tolksdorf R, Paslaru Bontas E. Engineering a semantic web for pathology. Proceedings of ICWE, 2004. 2004.

pp. 585–586. Available: http://swpatho.ag-nbi.de/papers/icwe04.pdfvia the INTERNET.

30. Aghbari Z, Makinouchi A. Semantic approach to image database classification and retrieval. NII Journal

2003;7:1–9. Available: http://www.nii.ac.jp/journal/pdf/07/07-01.pdf via the INTERNET.

31. Berners-Lee T, Hendler J, Lassila O. The semantic web. Scientific American 2001 May; Available: http://

www.sciam.com/2001/0501issue/0501ber ners-lee.html via the INTERNET.

32. Paslaru Bontas E, Tietz S, Schrade TR. Experiences using semantic web technologies to realize an information

retrieval system for pathology. Berliner, XML Tage; 2004. pp. 82–93.

33. ISO 13250. Guide to the topic map standards. ISO/IEC JTC1/SC34; 2002. Available: http://www1.y12.doe.gov/

capabilities/sgml/sc34/document/0323.htm.

34. ISO 13250.Topic Maps, 2nd ed. ISO/IEC JTC1/SC34; 2002. Available: http://www1.y12.doe.gov/capabilities/

sgml/sc34/document/0322_files/iso13250-2nd-ed-v2.pdf.

35. Garshol LM. Topic maps—ontologies for humans? XML UK, XML and ontologies. Available: http://

www.xmluk.org/images/content/Ontologies_and_XML_12_05/tms-ontologies-for-humans.pdf via the INTER-

NET. Accessed 2005 Dec 08.

36. Pepper S, Moore G, editors. XTM 1.0, XML Topic Maps (XTM) 2001. Available from: http://

www.topicmaps.org/xtm/1.0/.

37. Ahmed K. Topic maps for open (source) developers. XTech 2005: XML, the web and beyond 2005. Available:

http://www.idealliance.org/proceedings/xtech05/papers via the INTERNET. Accessed 2004 Mar.

38. Pepper S. The TAO of topic maps seamless knowledge in practice 2005; Ontopia AS. Available: http://

www.ontopia.net/topicmaps/materials/Towards%20Seamless%20Knowledge.ppt via the INTERNET.

39. Moore G. Topic Map technology—the state of the art 200. Available: http://www.infoloom.com/gcaconfs/WEB/

paris2000/S22-04.HTM via the INTERNET.

40. Vatant B. Ontology-driven topic maps 2004. Available: http://www.idealliance.org/europe/04/call/xmlpapers/03-

03-03.91/.03-03-03.html via the INTERNET.

41. Wrightson A. Topic maps and knowledge representation 2001. Available: http://www.ontopia.net/topicmaps/

materials/kr-tm.html via the INTERNET.

42. Ontopia Tolog. Available from: http://www.ontopia.net/topicmaps/materials/tolog.html.

43. Ontopia, Tolog tutorial, Ontopia AS Internet 2001. Available from: http://www.ontopia.net/omnigator/docs/

query/tutorial.htm.

44. Garshol LM. TOLOG, a topic map query language 2001; Available: http://www.ontopia.net/ via the

INTERNET.

45. Papastergiou A, Grammatikopoulos G, Hatzigaidas A, Tryfon G. Presenting a 3-tier system for managing

melanoma images with topic maps. WSEAS Trans Information Science and Applications 2007;4:820–826.

46. Netbeans IDE Organisation. Available from: http://www.netbeans.org.

47. AJAX: Getting started. Available from: http://developer.mozilla.org/en/docs/Main_Page.

48. AJAX: asynchronously moving forward. Available from: http://www.interaktonline.com/Support/Articles/Details/

AJAX:+Asynchronously+Moving+Forward-Why+use+AJAX?.html?id_art=36.

49. Asynchronous JavaScript technology and XML (AJAX) with the Java. Available from: http://java.sun.com/

developer/technicalArticles/J2EE/AJAX/.

50. Papastergiou A, Hatzigaidas, A, Tryfon G, Ioannidis D, Grammatikopoulos G, editors. Employing melanoma

diagnosis support using ajax and topic maps. Proceedings of the 11th WSEAS International Conference on

COMPUTERS (part of the 11th WSEAS CSCC Multiconference); 2007 Jul 26–28; Agios Nikolaos, Crete

Island, Greece, 2007. p. 318.


Dow

nloa

ded

By:

[HE

AL-

Link

Con

sorti

um] A

t: 13

:13

12 J

une

2008

51. MySQL. Available from: http://www.mysql.com/.

52. TM4J: Topic maps for Java. Available from: http://www.tm4j.org/.

53. Google Web Toolkit—Build AJAX apps in the Java language. Available from: http://code.google.com/webtoolkit/.

54. Papastergiou A, Hatzigaidas A, Grammatikopoulos G, Zaharis Z, Lazaridis P, Kampitaki D, Tryfon G.

Introducing an advanced Topic Map software tool towards the deployment of a TM-based system for managing

melanoma cases images. WSEAS Trans Information Science and Applications 2007;3:452–459.

55. Papastergiou A, Grammatikopoulos G, Hatzigaidas A, Zaharis Z, Lazaridis P, Kampitaki D, Tryfon G,

Ioannidis D. Applying topic map technology to create a knowledge base of melanoma images for diagnostic

support. WSEAS Trans Information Science and Applications 2007;2:406–412.


a web-based melanoma image diagnosis support … topic occurrence speciﬁes information, ......

Documents