Adaptive Multimedia Retrieval: User, Context, and Feedback (eBook)

Title 2
Preface 6
Organization 8
Table of Contents 10
A Method for Processing the Natural Language Query in Ontology-Based Image Retrieval System 13
SAFIRE: Towards Standardized Semantic Rich Image Annotation 24
Ontology-Supported Video Modeling and Retrieval 40
Learning to Retrieve Images from Text Queries with a Discriminative Model 54
A General Principled Method for Image Similarity Validation 69
Rank-Test Similarity Measure Between Video Segments for Local Descriptors 83
Can Humans Benefit from Music Information Retrieval? 94
A New Approach to Probabilistic Image Modeling with Multidimensional Hidden Markov Models 107
3D Face Recognition by Modeling the Arrangement of Concave and Convex Regions 120
Fuzzy Semantic Action and Color Characterization of Animation Movies in the Video Indexing Task Context 131
Retrieval of Document Images Based on Page Layout Similarity 148
Multimedia Content Adaptation Within the CAIN Framework Via Constraints Satisfaction and Optimization 161
Aspects of Adaptivity in P2P Information Retrieval 176
Interactive Museum Guide: Accurate Retrieval of Object Descriptions 191
Semantic Image Retrieval Using Region-Based Relevance Feedback 204
Image Retrieval with Segmentation-Based Query 219
Fast Structuring of Large Television Streams Using Program Guides 234
Variation of Relevance Assessments for Medical Image Retrieval 244
An Efficient Collaborative Information Retrieval System by Incorporating the User Profile 259
The Potential of User Feedback Through the Iterative Refining of Queries in an Image Retrieval System 270
Author Index 281

STRONG>1 Introduction (p. 11)
Nowadays, the study on the image retrieval has been actively progressing. Until now, the basic image retrieval methodologies are the Text-Matching, Contents-based and Concept(Ontology)-based methods.[2][3] In these methodologies, users generally use simple keywords as the user query. The Ontology-based image retrieval system uses the ontologies to understand the meaning of the user query, but the ontologies just solve the ambiguousness between words. Hence, the user query used in ontologybased system is also simple keywords. Nowadays, huge number of images has been creating through the various image acquisition devices such as the digital camera, scanner and phone-camera.

Thus, we need more intelligent image retrieval techniques for searching the images efficiently. In present day, the users tend to use a descriptive sentence to find images because they want to search for images as fast as possible, they do not want to spend long time retrieving images. Thus, the user query is getting descriptive and natural language type. As a result, the method for processing the natural language query is demanded for improving the performance of the image retrieval system. In this paper, we use two kinds of ontologies in our proposed system to handle the natural language query.

One is the domain ontology, which contains many concepts and represents the relations between these concepts. The other is the spatial ontology, which contains three basic relations and many words about the relations. We use some parts of the WordNet for building the domain ontology and we newly make the spatial ontology based on the survey paper, WordNet and OXFORD Dictionary for the purpose of processing the natural language queries. The basic idea of our study is that most user queries are including the words representing the spatial relationships. It is the significant feature of user queries for supporting our study.

Therefore we use the features to design the newly proposed image retrieval system and try to process the natural language queries. In the 2nd Section, we introduce the related works - the ontology-based image retrieval and the query processing methodologies. Then in Section 3, we explain the spatial ontology building steps and our system architecture based on the ontologies. And we describe the method for processing the natural language queries in the ontology- based system in details. We test and evaluate our system comparing with other systems in Section 4. At the end of this paper, we conclude our study and suggest the future works.

2 Related Works

2.1 Ontology-Based Image Retrieval
The traditional information retrieval systems have the mismatch problem among the terminologies. For solving the problem, many researchers have studied to apply the ontology theory to the system. Many works show that ontologies could be used not only for annotation and precise information retrieval, but also for helping the user in formulating the information need and the corresponding query.

It is important especially in applications where the domain semantics are complicated and not necessarily known to the user. Furthermore, the ontology-enriched knowledge base of image metadata can be applied to construct more meaningful answers to queries than just hit-lists. The major difficulty in the ontology-based approach is that the extra work is needed in creating the ontology and the detailed annotations.