Takashi Kitagawa

A metadatabase system for semantic image search by a mathematical model of meaning

In the design of multimedia database systems, one of the most important issues is to extract images dynamically according to the users impression and the images contents. In this paper, we present a metadatabase system which realizes the semantic associative search for images by giving keywords representing the users impression and the images contents.This metadatabase system provides several functions for performing the semantic associative search for images by using the metadata representing the features of images. These functions are realized by using our proposed mathematical model of meaning. The mathematical model of meaning is extended to compute specific meanings of keywords which are used for retrieving images unambiguously and dynamically. The main feature of this model is that the semantic associative search is performed in the orthogonal semantic space. This space is created for dynamically computing semantic equivalence or similarity between the metadata items of the images and keywords.

A mathematical model of meaning and its application to multidatabase systems

One of the most important issues in the multidatabase research is semantic heterogeneity in autonomous databases. The authors propose a new model for realizing semantic interoperability among data items in multidatabase systems. In multidatabase systems, it is not easy to select appropriate databases and extract significant information for users requests from many databases which are managed independently by its own database management system. One of the hardest problems is that it is difficult to judge equivalence or similarity between data items which are included in different databases. To select appropriate databases and extract significant information for users requests, they have designed a meta database system. In this system. a mathematical model of meaning is used to find different data items with the equivalent or similar meaning or to recognize the different meanings of a data item. The mathematical model of meaning consists of: (1) defining a normal space, (2) constructing a class of projections which represents a phase of meaning, (3) constructing a mechanism to select a subspace of the normed space according to the context. The main feature of this model is that the specific meaning of a data item can be recognized disambiguously and dynamically according to the context.<<ETX>>

A fundamental framework for realizing semantic interoperability in a multidatabase environment

In multidatabase research, the realization of semantic interoperability is the most important issue for resolving semantic heterogeneity between different databases. In this paper, we propose a fundamental framework for realizing semantic interoperability at the level of semantic relationships between data items in a multi database environment. We present a metadatabase system which extracts the significant information from different databases. The metadatabase system uses a mathematical model of meaning to dynamically recognize the semantic equivalence, similarity, and difference between data items. The essential feature of this model is that the specific meaning of a data item is dynamically fixed and unambiguously recognized according to the context by semantic interpretation mechanisms.

Regularization Using QR Factorization and the Estimation of the Optimal Parameter

In this paper we propose a direct regularization method using QR factorization for solving linear discrete ill-posed problems. The decomposition of the coefficient matrix requires less computational cost than the singular value decomposition which is usually used for Tikhonov regularization. This method requires a parameter which is similar to the regularization parameter of Tikhonovs method. In order to estimate the optimal parameter, we apply three well-known parameter choice methods for Tikhonov regularization.

An implementation method of associative search for heterogeneous mediadata utilizing the mathematical model of meaning and its application to image data and facial expression

This paper presents an implementation method of associative search for heterogeneous mediadata. This method applies several recognition mechanisms of human Kansei for realizing associative search. The feature of this method is to bridge over heterogeneous mediadata which exist independently as different database resources. In this paper, we show an associative search method with media-to-media transformation from a picture to a facial expression. We clarify the effectiveness of our method by showing several experimental results.

A semantic associative search method for WWW information resources

In the current World Wide Web (WWW) environment, it is important to realize intelligent and effective information acquisition mechanisms. We propose a semantic associative search method based on our mathematical model of meaning to realize intelligent and effective information acquisition. Our method provides a dynamic context recognition mechanism for information acquisition according to users queries given as contexts. We integrate two application systems of our semantic associative search method to realize an intelligent and effective information acquisition environment for WWW information resources. We apply our method to information retrieval on WWW information resources, which are identified by Uniform Resource Locators (URLs). This application system makes it possible to dynamically obtain semantically related information resources on WWW, according to users queries given as contexts. Furthermore, we have applied our method to queries and information resources described in multiple languages. We integrate those applications for supporting WWW information acquisition.

Automatic media content creation system according to an impression by recognition-creation operators

In this paper, we introduce an automatic media content creation system according to an impression. The system creates media content such as music, an image, etc. according to an impression given by a user. The impression is represented in words. Generally, there are two systems related to an impression. The first is a recognition system. The recognition system evokes an impression from media content. The other is a creation system. The creation system creates media content according to an impression. We construct a cycle of the impression via the recognition creation systems. In this cycle, it is mutual mapping between impressions and media contents. In this paper, we construct a recognition operator and a creation operator for realizing the impression cycle. Each operator is a relation of inverse calculation. We apply this system to music media. We introduce an automatic music creation system.

A construction method of a retrieval space by an evaluation of word distributions in documents

This paper presents a new construction method of a retrieval space by an evaluation of word distributions in documents. In order to realize a semantic associative search, it is necessary to construct a retrieval space which enables measurement of relations between each word. A creation method of retrieval space has been proposed that creates a data matrix by using a dictionary or a term dictionary. However, it is difficult to create a retrieval space without a dictionary or a term dictionary. This method can construct a retrieval space which enables measurement of relations between each word by an evaluation of word distributions in documents. This paper shows some experimental results which applied the mathematical model of meaning to the retrieval space constructed by this method.

A construction method of a metadata space based on relations between words from an index of a book

A large amount of information resources for a specific field have been distributed in wide area networks. One of the most important issues is how to extract appropriate information in such environment for a specific field. This paper presents a construction method of a metadata space from an index of a book for a specific field. In order to search high quality information on a specific field, it is necessary to construct a metadata space which enables measurement of relations between each expert word for the field. This method needs neither a dictionary nor specialized knowledge. This method enables construction of a metadata space for an expert field easily.

The Method of Image Feature Selection for Integration of Image Classification by Bag-of-Keypoints

In this paper, we propose a method of image feature selection for integration of image classification by Bag-of-Keypoints method and efficient search method. Our method is integration of image classification, which provides high-precision classification for various images. Therefore, in order to select a type of image feature in accordance with various purposes, it is necessary to select method easily. In addition, our method integrates various kinds of image feature by Bag-of-Keypoints method. Classification with a combination of some detectors and descriptors is more effective than with single detector and descriptor. To realize combination of some detectors and descriptors, we propose an integrating method. This paper is LATE BREAKING PAPERS for CSCI-ISAI.