Sunkyoung Baek

Object retrieval by query with sensibility based on the KANSEI-Vocabulary scale

Recently the demand for image retrieval and recognizable extraction corresponding to KANSEI (sensibility) has been increasing, and the studies focused on establishing those KANSEI-based systems have been progressing more than ever. In addition, the attempt to understand, measure and evaluate, and apply KANSEI to situational design or products will be required more and more in the future. Particularly, study of KANSEI-based image retrieval tools have especially been in the spotlight. So many investigators give a trial of using KANSEI for image retrieval. However, the research in this area is still under its primary stage because it is difficult to process higher-level contents as emotion or KANSEI of human. To solve this problem, we suggest the KANSEI-Vocabulary Scale by associating human sensibilities with shapes among visual information. And we construct the object retrieval system for evaluation of KANSEI-Vocabulary Scale by shape. In our evaluation results, we are able to retrieve object images with the most appropriate shape in term of the querys KANSEI. Furthermore, the method achieves an average rate of 71% users satisfaction.

Matching colors with KANSEI vocabulary using similarity measure based on wordnet

Recently, the image retrieval based on content is capable of understanding the semantics of visual information. However, it is hard to represent emotion or feeling of human. To approach more intelligent content-based retrieval, we focus on KANSEI information. This paper presents a method of matching color, which is part of visual information associated with KANSEI-vocabulary relation. We use WordNet that is a kind of lexical ontology by relations between words. We define relation for matching between color and KANSEI vocabulary using the meaning of color table. We propose the similarity measure between Color-KANSEI vocabulary and query. After experiment we can find the best pertinent color using Lesk algorithm. The significance of our study is finding semantically pertinent color according to various queries based on WordNet. This is the approach as computing vocabulary to show KANSEI of Human.

Grasping related words of unknown word for automatic extension of lexical dictionary

An aim of this research is to grasp related words of unknown word. Currently, several lexical dictionaries have been developed for semantic retrieval such as WordNet and FrameNet. However, more new words are created in every day because of new trends, new paradigm, new technology, etc. And, it is impossible to contain all of these new words. The existing methods, which grasp the meaning of unknown word, have a limitation that is not exact. To solve this limitation, we have studied the way how to make relations between known words and unknown word. As a result, we found a noble method using co-occurrence, wordnet and Bayesian probability. The method could find what words are related with unknown word and how much weight other words relate with unknown word.

KANSEI-Based image retrieval associated with color

Nowadays, the processing of KANSEI information is very important in intelligent computing field. Particularly, it is very interesting in image retrieval to deal with humans KANSEI. In this paper, we use natural language for the representation of KANSEI, including the image structure of Humans idea, which we can not observe. And then, a KANSEI-Adjective is used as a natural language querying method: In other words, this paper presents the image retrieval based on KANSEI. We propose the background image retrieval based on KAC (KANSEI-Adjective of Color) to represent the sensibility of color. Our method for processing of KANSEI information is the measure of similarity by using the adaptive Lesk algorithm in WordNet. In our experimental results, we are able to retrieve background images with the most appropriate color in term of the querys feeling. Furthermore, the method achieves an average rate of 63% users satisfaction.

The Techniques for the Ontology-Based Information Retrieval

The use of ontologies to address the problems of the existing keyword-based search has been searched. For the efficient ontology-based information retrieval, there are several facts we should consider. In this paper, we describe the techniques demanded for the ontology-based information retrieval.

Adaptive-Tangent space representation for image retrieval based on kansei

From the engineering aspect, the research on Kansei information is a field aimed at processing and understanding how human intelligence processes subjective information or ambiguous sensibility and how such information can be executed by a computer. Our study presents a method of image processing aimed at accurate image retrieval based on human Kansei. We created the Kansei-Vocabulary Scale by associating Kansei of high-level information with shapes among low-level features of an image and constructed the object retrieval system using Kansei-Vocabulary Scale. In the experimental process, we put forward an adaptive method of measuring similarity that is appropriate for Kansei-based image retrieval. We call it “adaptive-Tangent Space Representation (adaptive-TSR)”. The method is based on the improvement of the TSR in 2-dimensional space for Kansei-based retrieval. We then it define an adaptive similarity algorithm and apply to the Kansei-based image retrieval. As a result, we could get more promising results than the existing method in terms of human Kansei.

A method for processing the natural language query in ontology-based image retrieval system

There is a large amount of image data on the web because of the development of many image acquisition devices nowadays. Hence, many researchers have been focusing on the study how to manage and retrieve these huge images efficiently. In this paper, we use two kinds of ontologies in the image retrieval system for processing the natural language query. We use the domain ontology for describing objects in images and we newly build the spatial ontology for representing the relations between these objects. And then, we suggest the method for processing the user query formatted by the natural language in the ontology-based image retrieval system. Based on our study, we got the conclusion that the natural language query processing is the very important part for improving the efficiency of the image retrieval system.

Measurement of arc-value for concept similarity

The ultimate aim of this paper is for semantic information retrieval. Especially, we concentrate on measure of concept similarity. There are 9 kinds of relation (arc) in WordNet to define concepts and we have a motive from that a value of each arc is different up to kind of relation. So, this paper first makes a fundamental for the value of relation between concepts. And we measure the arc-value based on the fundamental and calculate concept similarity using the value.

Kansei Processing Agent for Personalizing Retrieval

In the present, methods of creating and processing a profile are insufficient for achieving personalization information retrieval that reflects the subjective Kansei preference of users. To rectify this insufficiency, we have created a Kansei information processing agent. Our study proposes a Kansei agent for the creation, accumulation and renewal of profiles in personalized retrieval and explores possible contributions to the development of a Kansei-based recommendation system and personalizing service.

Semantic Image Analysis Based on the Representation of the Spatial Relations Between Objects in Images

The number of images available on the world wide web has grown enormously, because of the increasing use of scanners, digital cameras and camera-phones. Consequently, the efficient retrieval of images from the web is necessary. Most existing image retrieval systems are based on the text or content associated with the image. In this paper, we propose a semantic image analysis for the semantic web. We use the description about the image and try to represent it using OWL. We also define new axioms for representing the spatial relationships based on the spatial description logics.