Tetsuya Saito’s research while affiliated with University of Tsukuba and other places

According to data from Statistical Research on Shopping Centers (SC), the number of SC in January 2012 was 3050, which is 1.41 times that of Y2000. Despite gathering the impact on the custom absorbency power, there are some spaces with a negative atmosphere where crowds of people gather together at the same time in what could be considered as inactive dead spaces. In this paper, we propose two entertainment systems which display animations by object detection to improve these negative atmospheres of particularly large SCs. We survey the changes in customer impressions by placing these systems in SC and verify that these systems deliver better impressions on the spaces.

Currently interactive entertainment characters have their conversation topics prepared beforehand by the creator. In this paper, we propose the automatic chat generation engine for emotional entertainment characters using real-time news information. The character emotionally reacts to the news contents based on their interest and feelings so that the users can have intimate feeling. The character also incrementally learns user's interest from their response.

In traditional English learning as a second language, learners rarely have the opportunity to practice oral communication, so the acquisition of oral proficiency is a slow process. In this paper, we propose a task-based second language learning game system. The task-based learning method enables learners to obtain communicative skills through the practice of particular "missions" using voice and gesture communications with life-size 3D game character.

In games and movies that use characters, utterances are important expressions of action. However, currently, interactive characters have their conversation topics prepared beforehand by the creator. If the topics are few, they repeat the same content with fixed variations, resulting in the user losing interest. In this paper, we propose the concept of entertainment character for emotional information display. We use RSS, which provides live feeds on news and weather forecasts on the Internet, to provide real-world information thereby enriching characters conversation topics. And, we generate various conversations between the user and the characters by adding topics related to the utterances and feelings appropriate to the topic for improving the expressiveness of talk.