Current Students


PhD Students


Hilal Tarakçı        Multimedia and semantic web

Ali Anıl Sınacı

Tolga Can



MS Students


Gizem Öztürk       Personalization in multimedia retrieval

Çağla Yaman       Ontology-based Personalization in video databases

Müjdat Bayar       Video Analysis for event detection

Esra Abacıoğlu     Automatic object recognition in videos

Goncagül Demirdizen    Ontology-based multimedia database systems

Gökhan Yaprakkaya     Automatic Face Rcognition in videos






Former Students


1.       Yakup Yildirim    Formalizing Workflows Using the Event Calculus, 2000



The event calculus is a logic programming formalism for representing events and their effects especially in database applications. This thesis presents the use of the event calculus for specifying and simulating workflows. The proposed framework maintains a representation of the dynamic world being modeled on the basis of user supplied axioms about preconditions and effects of events and the initial state of the world. The net effect is that a workflow specification can be made at a higher  level of abstraction. Within this framework it is possible to model sequential and concurrent activities with synchronization when necessary. It is also possible to model agent assignment and concurrent workflow instances. A logic programming approach to the computational problem is adopted.


2.       Mesru Koprulu    Spatio-temporal querying in video databases, 2001



In this thesis a video data model that allows efficient and effective representation and querying of spatio-temporal properties of objects is presented. The data model is focused on the semantic content of video streams. Objects, events, activities performed by objects are main interests of the model. The model supports fuzzy spatial queries including querying spatial relationships between objects and querying the trajectories of objects. The model is flexible enough to define new spatial relationship types between objects without changing the basic data model. A prototype of the proposed model has been implemented. The prototype allows various spatio-temporal queries along with the fuzzy ones and it is prone to implement compound queries without major changes in the data model.


3.       Esen Kacar           Mining Fuzzy Spatial Association Rules, 2001



Discovering interesting, implicit knowledge and general relationships in geographic information databases is very important to understand and to use these spatial data. One of the methods for discovering this implicit knowledge is mining spatial association rules. A spatial association rule is a rule indicating certain association relationships among a set of spatial and possibly non-spatial predicates. In the mining process, data is organized in a hierarchical manner. However, in real-world applications it may not be possible to construct a crisp structure for this data, rather some fuzzy structures should be used. Fuzziness, i.e. partial belonging of an item to more than one sub-item in the hierarchy, could be applied to the data itself, and also to the hierarchy of spatial relations. This thesis shows that, strong association rules can be mined from large spatial databases using fuzzy concept and spatial relation hierarchies.


4.       Narin Emeksiz      Diagnosis of Gastric Carcinoma Tumours by Multi Class Voting, 2001


5.       Deniz Canturk      Time-based workflow mining, 2005



Creating a workflow design is a complicated process and usually there are differences between the actual workflow processes and the processes as seen by the designer. Therefore, new data mining techniques for (re)discovering workflow models have been proposed. Starting point for such techniques are so-called “workflow logs" which contains information about the workflow process as it is actually being executed. In this paper, a new workflow mining technique based on time information is proposed. It is assumed that events in workflow logs are recorded with their start and end timestamps. This information is used in order to determine task orders and control flows between tasks. This new algorithm can be used to mine workflow graphs which contain basic workflow structures, such as sequential, parallel, iterative routings and or-join. The proposed algorithm finds a concrete workflow model as an output. In this workflow model execution durations of tasks and waiting times between tasks are included; hence the start time of each task and total execution time of the workflow can be discovered. This information could also be used in commenting on hidden tasks.


6.       Guzen Erozel        Natural Language Interface on a video data model, 2005



The video databases and retrieval of data from these databases have become popular in various business areas of work with the improvements in technology. As a kind of video database, video archive systems need user-friendly interfaces to retrieve video frames. In this thesis, an NLP based user interface to a video database system is developed using a content-based spatio-temporal video data model. The data model is focused on the semantic content which includes objects, activities, and spatial properties of objects. Spatio-temporal relationships between video objects and also trajectories of moving objects can be queried with this data model. In this video database system, NL interface enables flexible querying. The queries, which are given as English sentences, are parsed using Link Parser. Not only exact matches but similar objects and activities are also returned from the database with the help of the conceptual ontology module to return all related frames to the user. This module is implemented using a distance-based method of semantic similarity search on the semantic domain-independent ontology, WordNet. The semantic representations of the given queries are extracted from their syntactic structures using information extraction techniques. The extracted semantic representations are used to call the related parts of the underlying spatio-temporal video data model to calculate the results of the queries.


7.       Onur Aydin                   Automated web service composition with event calculus, 2005



As the Web Services proliferate and complicate it is becoming an overwhelming job to manually prepare the Web Service Compositions which describe the communication and integration between Web Services. This thesis analyzes the usage of Event Calculus, which is one of the logical action-effect definition languages, for the automated preparation and execution of Web Service Compositions. In this context, planning capabilities of Event Calculus are utilized. Translations from Planning Domain Description Language and DARPA Agent Markup Language to Event Calculus are provided to show that Web Services can be composed with the help of Event Calculus. Also comparisons between Event Calculus and other planning languages used for the same purposes are presented.


8.       Cigdem Celik        An MPEG-7 Video Database System for content-basedf management and retrieval, 2005



A video data model that allows efficient and effective representation and querying of spatio-temporal properties of objects has been previously developed. The data model is focused on the semantic content of video streams. Objects, events, activities performed by objects are the main interests of the model. The model supports fuzzy spatial queries including querying spatial relationships between objects and querying the trajectories of objects. In this thesis, this work is used as a basis for the development of an XML-based video database system. This system is aimed to be compliant with the MPEG-7 Multimedia Description Schemes in order to obey a universal standard. The system is implemented using a native XML database management system. Query entrance facilities are enhanced via integrating an NLP interface.


9.       Funda Karagoz     Application of Schema Matching Methods to Semantic Web Service Discovery, 2006



The Web turns out to be a collection of services that interoperate through the Internet. As the number of services increase, it is getting more and more diffucult for users to find, filter and integrate these services depending on their requirements. Automatic techniques are being developed to fulfill these tasks. The first step toward automatic composition is the discovery of services needed. UDDI which is one of the accepted web standards, provides a registry of web services. However representation capabilities of UDDI are insufficient to search for services on the basis of what they provide. Semantic web initiatives like OWL and OWL-S are promising for locating exact services based on their capabilities. In this thesis, a new semantic service discovery mechanism is implemented based on OWL-S service profiles. The service profiles of an advertisement and a request are matched based on OWL ontologies describing them. In contrast to previous work on the subject, the ontologies of the advertisement and the request are not assumed to be same. In case they are different, schema matching algorithms are applied. Schema matching algorithms find the mappings between the given schema models. A hybrid combination of semantic, syntactic and structural schema matching algorithms are applied to match ontologies.


10.     Huseyin Yilmaz    Realizing the specification and execution of workflows through event calculus, 2006



Workflow management promises a solution to an age-old problem: controlling, monitoring, optimizing and supporting business processes. What is new about workflow management is the explicit representation of the business process logic which allows for computerized support. In the light of this support, many researchers developed different approaches to model new systems with different capabilities to solve this age-old problem. One of the approaches is using logic-based methodology for the specification and execution of workflows. Here, the event calculus, a logic programming formalism for representing events and their effects especially in database applications, is used for this approach. It is shown that the control flow graph of a workflow specification can be expressed as a set of logical formulas and the event calculus can be used to specify the role of a workflow manager through a set of rules for the execution dependencies of activities. Constructed workflow formalization through Event Calculus is realized by using recent technologies, and the resulting product is named as EventFlow, including some administrative interfaces to manage system and workflow engine. The architecture and implementation details of EventFlow, an editor developed for graphical representation of control flow graph, and used technologies are described. And an example application is built to show the usability and execution of the implemented system.


11. Esra Kırcı       Automatic Composition of Semantic Web Services with the Abductive Event Calculus, 2007



This thesis investigates the application of AI planning techniques to the web service composition problem and in particular, it proposes the use of the abductive event calculus in this domain. Web service compositions are defined as templates using OWL-S ("OWL for Services"). These generic composition definitions are converted to Prolog language as axioms for the abductive event calculus planner and solutions found by the planner constitute the specific result plans for the generic composition plan. In this thesis it is shown that abductive planning capabilities of the event calculus can be used to generate the web service composition plans that realize the generic procedure.


12. Esat Kaan Kuban    Abductive Planning Approach for Automated Web Service Composition using only User Specified Inputs and Outputs, 2008



Automated web service composition can be achieved either by generating the composition plan dynamically using given inputs and outputs, or by locating the correct services if an abstract process model

is given. This thesis investigates the former method which is dynamicly generating the composition by using the abductive planning capabilities of the Event Calculus. Event calculus axioms in Prolog language, are generated using the available OWL-S web service descriptions in the service repository, values given to selected inputs from ontologies used by those semantic web services and desired output types selected again from the ontologies. Abductive Theorem Prover which is the AI planner used in this thesis, generates composition plans and execution results according to the generated event calculus axioms.





In this thesis, we present an approach of metadata extraction from match reports for soccer domain. The UEFA Cup and UEFA Champions League Match Reports are downloaded from the web site of UEFA by a web-crawler. These match reports are preprocessed by using regular expressions and then important events are extracted by using hand-written rules. In addition to  hand-written rules, two different machine learning techniques are applied on match corpus to learn event patterns and automatically extract match events. Extracted events are saved in an MPEG-7 file. A user interface is implemented to query the events in the MPEG-7 match corpus and view the corresponding video segments.





In this thesis, MPEG-7 ontology is used on top and the given user-defined ontologies are attached to the MPEG-7 ontology via a user friendly interface, thus building MPEG-7 based ontologies automatically. Our proposed system is an ontology-based multimedia information management framework due to its modular architecture, ease of integrating with domain specific ontologies naturally and automatic harmonization of MPEG-7 ontology and domain-specific ontologies. Integration with domain specific ontologies is carried out by enabling import of domain ontologies via a user-friendly interface which makes the system independent of application domains.





In this thesis, some novel ideas are presented for solving automated web service composition problem. Some possible real world problems such as partial observability of environment, nondeterministic effects of web services, service execution failures are solved through some mechanisms. In addition to automated web service composition, automated web service invocation task is handled in this thesis by using reflection mechanism. The proposed approach is based on AI planning. Web service composition problem is translated to AI planning problem and a novel AI planner namely “Simplanner” that is designed for working in highly dynamic environments under time constraints is adapted to the proposed system. World altering service calls are done by conforming to the WS-Coordination and WS-Business Activity web service transaction specifications in order to physically repair failure situations and prevent undesired side effects of aborted web service composition efforts.


16. Onur Özorhan          A Service Oriented Peer to Peer Web Service Discovery Mechanism With Categorization



This thesis, studies automated methods to achieve web service advertisement and discovery, and presents efficient search and matching techniques based on OWL-S. In the proposed system, the service discovery and matchmaking is performed via a centralized peer-to-peer web service repository. The repository has the ability to run on a software cloud, which improves the availability and scalability of the service discovery. The service advertisement is done semi-automatically on the client side, with an automatic WSDL to OWL-S conversion, and manual service description annotation. An OWL-S based unified ontology -Suggested Upper Merged Ontology is used during annotation, to enhance semantic matching abilities of the system. The service advertisement and availability are continuously monitored on the client side to improve the accuracy of the query results. User-agents generate query specification using the system ontology, to provide semantic unification between the client and the system during service discovery. Query matching is performed via complex Hilbert Spaces composed of conceptual planes and categorical similarities for each web service. User preferences following the service queries are monitored and used to improve the service match scores in the long run.


17. Çağla Okutan           A Monolithic Approach to Automated Composition of Semantic Web Services with the Event Calculus



In this thesis, a web service composition and execution framework is presented for semantically annotated web services. A monolithic approach to automated web service composition and execution problem is chosen, which provides some benefits by separating the composition and execution phases. An AI planning method using a logical formalism called Event Calculus is chosen for the composition phase. This formalism allows one to generate a narrative of actions and temporal orderings using abductive planning techniques given a goal. Functional properties of services, namely input/output/precondition/effects(IOPE) are taken into consideration in the composition phase and non-functional properties, namely quality of service (QoS) parameters are used in selecting the most appropriate solution to be executed. The repository of OWL-S semanticWeb services are translated to Event Calculus axioms and the resulting plans found by the Abductive Event Calculus Planner are converted to graphs. These graphs can be sorted according to a score calculated using the defined quality of service parameters of the atomic services in the composition to determine the optimal solution. The selected graph is converted to an OWL-S file which is executed consequently.


18. Atakan Şimşek        Ontology-Based Spatio-Temporal Video Management System



In this thesis, a system, called Ontology-Based Spatio-Temporal Video Management System (OntoVMS) is developed in order to supply a framework which can be used for semantic data modeling and querying in video files. OntoVMS supports semantic data modeling which can be divided into concept modeling, spatio-temporal relation and trajectory data modeling. The system uses Rhizomik MPEG-7 Ontology as the core ontology. Moreover ontology expression capability is extended by automatically attaching domain ontologies. OntoVMS supports querying of all spatial relations such as directional relations (north, south ...), mixed directional relations (northeast, southwest ...), distance relations (near, far), positional relations (above, below ...) and topological relations (inside, touch ...); temporal relations such as starts, equal, precedes; and trajectories of objects of interest. In order to enhance querying capability, compound queries are added to the framework so that the user can combine simple

queries by using “(”, “)”, “AND” and “OR” operators. Finally, the use of the system is demonstrated with a semi-automatic face annotation tool.


19. Mennan Güder         Data Mining Methods For Clustering Power Quality Data Collected Via Monitoring Systems Installed On The Electricity Network



Increasing power demand and wide use of high technology power electronic devices result in need for power quality monitoring. The quality of electric power in both transmission and distribution systems should be analyzed in order to sustain power system reliability and continuity. This analysis is possible by examination of data collected by power quality monitoring systems. In order to define the characteristics of the power system and reveal the relations between the power quality events, huge amount of data should be processed. In this thesis, clustering methods for power quality events are developed using exclusive and overlapping clustering models. The methods are designed to cluster huge amount of power quality data which is obtained from the online monitoring of the Turkish Electricity Transmission System. The main issues considered in the design of the clustering methods are the amount of the data, efficiency of the designed algorithm and queries that should be supplied to the domain experts. This research work is fully supported by the Public Research grant Committee (KAMAG) of TUBITAK within the scope of National Power quality Project (105G129).


20. Kezban Demirtaş     Automatic Video Categorization And Summarizatıon



In this thesis, we make automatic video categorization and summarization by using subtitles of videos. We propose two methods for video categorization. The first method makes unsupervised categorization by applying natural language processing techniques on video subtitles and uses the WordNet lexical database and WordNet domains. The method starts with text preprocessing. Then a  keyword extraction algorithm and a word sense disambiguation method are applied. The WordNet domains that correspond to the correct senses of keywords are extracted. Video is assigned a category label based on the extracted domains. The second method has the same steps for extracting WordNet domains of video but makes categorization by using a learning module. Experiments with documentary videos give promising results in discovering the correct categories of videos. Video summarization algorithms present condensed versions of a full length video by identifying the most significant parts of the video. We propose a video summarization method using the subtitles of videos and text summarization techniques. We identify significant sentences in the subtitles of a video by using text summarization techniques and then we compose a video summary by finding the video parts corresponding to these summary sentences.


21. Erkin Eryol              Probabilistic Latent Semantic Analysis Based Framework For Hybrid Social Recommender Systems



Today, there are user annotated internet sites, user interaction logs, online user communities which are valuable sources of information concerning the personalized recommendation problem. In the literature, hybrid social recommender systems have been proposed to reduce the sparsity of the usage data by integrating the user related information sources together. In this thesis, a method based on probabilistic latent semantic analysis is used as a framework for a hybrid social recommendation system. Different data hybridization approaches on probabilistic latent semantic analysis are experimented. Based on this flexible probabilistic model, network regularization and model blending approaches are applied on probabilistic latent semantic analysis model as a solution for social trust network usage throughout the collaborative filtering process. The proposed model has outperformed the baseline methods in our experiments. As a result of the research, it is shown that the proposed methods successfully model the rating and social trust data together in a theoretically principled way.



22. Soner Kara              An Ontology-Based Retrieval System Using Semantic Indexing



In this thesis, we present an ontology-based information extraction and retrieval system and its application to soccer domain. In general, we deal with three issues in semantic search, namely, usability, scalability and retrieval performance. We propose a keyword-based semantic retrieval approach. The performance of the system is improved considerably using domain-specific information extraction, inference and rules. Scalability is achieved by adapting a semantic indexing approach. The system is implemented using the state-of-the-art technologies in SemanticWeb and its performance is evaluated against traditional systems as well as the query expansion methods. Furthermore, a detailed evaluation is provided to observe the performance gain due to domain-specific information extraction and inference. Finally, we show how we use semantic indexing to solve simple structural ambiguities.