Clustered table or matrix data is often very important in various computer science fields. Many of recently reported vsiualization techniques for the table data convert them into hierarchical or graph data. This paper presents "Sakyo & Ukyo", a visualization technique for table or matrix data, applying a hierarchical data visualization technique "HeiankyoView". The technique first applies clustering for rows and columns of table data independently, and constructs two hierarchical data from the clustering results. It then independently applies "HeiankyoView" to the two hierarchical data. It realizes a new visualization schema by providing a mechanism to interact the two hierarchical data each other. We have applied "Sakyo & Ukyo" to the visualization of table data constructed from the corpus of newspaper articles. This paper introduces that we discovered intersting clusters of keywords and documents by "Sakyo & Ukyo".

View full abstract

This paper presents an interactive system which controls fountains. The system consists of seven fountain units; each unit has a PC-controlled water pump and nine full-color LED lights.Each fountain unit pump is is covered with a clear acrylic pipe. The lighting color and jet speed of water are controlled with motion of player′s fan.We have installed an interactive fountain system applying an affordance of control fan.

View full abstract

This paper propose a novel visualization system for exploration of recipes. The goal of the system is to support users in exploring a large amount of recipe data for the purpose of decision-making. The system visualizes relation between recipes and ingredients as graph. User can expand the displayed networks by only selecting an ingredient.

View full abstract

This paper proposes a system to visualize the historical data and relations between timelines. The system gives user a clear overview that user couldn't get from the usual text-based linear list of events. The system has two main visualization features as follows. 1) Distributions of events within a single timeline are represented using a timeline object. The vertical axis corresponds to a timeline, and the thickness of the object represents the number of events that occurs at a certain time. 2) Users input a keyword, then the keyword represented by a dot. The system retrieves all event data contained in the keyword string. When the system finds the event data containing the keyword, a line connects the keyword to the timeline it matches. Therefore, two timelines can be connected through a common keyword from their respective events. We have conducted an experiment to investigate the effects of our system.

View full abstract

Abstract-This paper proposes a virtual fighting game player who speaks on various situations from game progress. The method of capturing a game status of Life-point-gauge and skill-taking-patterns is read from special pixels on game video graphic frames. Emotional parameters are calculated from this value. Utterance sound is captured and classified as vice-set at on same time. After those preparations, Game is played with this system. Appropriate voices made from real speech on playing person is selected voice-set by situation. A user evaluation shows this method adds game joy increased.

View full abstract

In recent years, improvisational musical expression systems have gained much attention.The authors particularly focus on systems that allow users both to realize their creativity freely and to practice musical techniques. In this article, we define the contexts of such systems and explore effective supporting methods by looking closely at various contemporary systems. The future developments of these systems′ design are also discussed.

View full abstract

This paper presents a technique for visualizing large-scale hierarchical multi-variate data, JunihitoeView. The technique is an extension of HeiankyoView, an existing technique for visualizing large-scale hierarcical data.HeiankyoView represents data items as icons, but JunihitoeView represents them by assigning various colors to represent the values of the multi-variate by their brightness.The paper also presents a level-of-detail control techniquefor JunihitoeView.The technique displays variables of a leaf node themselves,when a user zooms in the visualization result. Further, it unifies the variables and display typical values of a higher level hierarchy,when the user zooms out the visualization result.The paper discusses the usefulness of JunihitoeView with the application of the technique for the hierarchicalmulti-variate data constructed from a drug database.It also shows results of user tests of the presented techniques.

View full abstract