In recent years, Japanese film’s revenue has more increase than before. For gaining a lot of films’ performance income, it is important for the customer to occur the unusual emotion. It is usually said that the succeeding movie has been introducing a various emotional tendency in the scene and customer’s unusual emotional undulations. A movie script must help the implementation method for creating the emotion. This research, therefore, focuses on the relationship among the movie script and the performance income. This research proceeds two ways: one is the movie script is counted and quantified by basic data of movies and cleared the relationship among its quantification and the movie’s revenue by statistical analysis. Second, basic information, such as an award history of directors and the number of performers, and evaluation of the box office revenue. An assessment framework sets and makes clear the relationship among the movie script, the award history, and the box office revenue. This research is finding two kinds of things: 1) It is able to reduce errors by using movie script and the emotional undulations relationship for predicting the box office revenue. 2) Young actors are useful to promote the main actors in romantic movies.

In order to convey user’s feeling through his/her avatar on remote communication, not only verbal information but also non-verbal information such as tone of voice, facial expression, and gesture are important. Although some studies propose methods to express basic emotions on avatar, no method has been established to create facial animation that expresses complicated emotional states like “happy but embarrassed” yet. This paper proposes a method to create facial animation expressing complicated emotional state by combining involuntary facial actions of suppressed real emotion and voluntary facial actions of suppressing fake emotion. The proposed method is explained in a case of “Tsundere” that consists of real happiness and fake calm/anger. 18 facial animations were created and evaluated by Sheffé’s method. The experimental results revealed that blushing, an involuntary action for happy, enhanced impression that the avatar sincerely felt happy. On the contrary, smiling faces without involuntary blushing gave impression that the avatar pretended to be happy with fake smile. It indicates that the proposed method that expresses real emotion by involuntary action unit can express real emotion selectively.

This paper proposes an interactive comic generation support system with effect materials. In the proposed system, short text with a story and images related to the topic are input, and a comic is interactively generated. It is an interactive system that allows users to select unsatisfied frames and regenerate only those frames after generating comic. It is also possible to regenerate in smaller parts, such as backgrounds, characters, balloons, drawn characters, and comic marks, in the selected frames. This makes it possible to quickly generate comics that are highly satisfying to users. Furthermore, the use of effect materials such as drawn characters, comic marks, and effect lines has been introduced. As a result, more comic-like expressions are possible. Comic is a more effective information transmission method than documents. Through the evaluation experiment, the effect of interactive regeneration and the usefulness by introducing drawn characters, comic marks, and effect lines have been confirmed.

In order to provide user-friendly recommendations, it is necessary to understand the user’s intentions and circumstances, and it is also essential to empathize based on non-verbal communication. In this paper, we develop a system that estimates the current state and ideals of users in health promotion and provides information support while empathizing. The purpose of the system is to provide information support according to the estimation of the user’s intentions and circumstances by introducing the following systems: a dialogue system with sympathy elements by multi-robot communication, and a reliability estimation system with user reactions as input. Furthermore, in order to verify the effectiveness of the proposed method, we show comparative experimental examples between the proposed method and the conventional method.

The effect of music on human emotion has been studied for a long time. Research on emotions for music, for example the research on such as feelings and impressions when listening to music, has been established as one research field. However, although there were many studies that cause an emotion from a music, few researches on creating a music from an emotion have been performed.

Therefore, in this study, we focus on facial expressions as emotional representation and aim to create a music that matches the emotion recognized from a facial image. For example, the system, which generates a bright and pleasant music using a laughing face image, or a dark and sad music using a crying face image automatically, will be constructed. Russell’s circumplex model was used for emotion recognition, and Hevner’s circular scale was used to generate music corresponding to these emotions. By using this system, for example, it will become possible to create a suitable BGM for the scene with only the actor’s face image in the production of movies. In this study, the above-mentioned system was constructed and the efficiency of this system was confirmed by conducting the Kansei evaluation experiment.

This paper focuses on a shadow box, a type of papercraft, and implements an impression-based fabrication support system that automatically generates shadow box expressions suitable for the user’s desired impression. The system consists of a design generation unit creating design candidates and an impression change evaluation model assessing the candidates in terms of their appropriateness to the desired impression. In developing the system, an impression evaluation experiment is first conducted to clarify the relationship between various shadow box expressions and their impression. By training neural networks using the experimental result, a set of impression change evaluation models is implemented. The design generation unit is built based on the approach of an evolutionary computation in which the implemented model set is used as an evaluation function. An impression evaluation experiment for the shadow boxes based on the system output demonstrates the validity of this system.

The possibility of rapid spread of highly infectious viruses necessitates prompt response. Being able to predict the effectiveness of particular measures against spread is therefore desirable. Although prediction of the transmission and extent of infection using multi-agent simulation is being carried out to this end, conventional simulation models do not incorporate the concept of isolation wards. Since isolation wards play roles in the quarantine and treatment of patients, they constitute an important parameter that affects fatality rate and transmission of infection. Prediction, therefore, that takes this point into consideration is advantageous. Another parameter that affects transmission is the extent by which people go outdoors. Particularly in Japan, a stay-at-home order was issued across wide areas around the country as a measure against COVID-19, which spread rapidly in Japan around April 2020. It is also important to be able to verify the effectives of these kinds of measures before they are implemented. Therefore, in this study, we developed a multi-agent simulation model for viral infection that incorporates the use of isolation wards and the implementation of stay-at-home orders. This simulator will enable the verification of the effectiveness of isolation wards and staying indoors in accordance with the infectiousness of the virus and the scale of the artificial society. This paper discusses several simulation patterns and their effectiveness.