Media contents are naturally used in our daily life, and cross-modal effects of them are believed. The main example of the cross-modal effects investigated in previous studies was a combination of sight and hearing. This study aims to fundamentally investigate the cross-modal effects of music and scent on subjective feelings and preference. Especially, the main effect of scent is focused on, because it has not been observed. Two music pieces and two scents were employed and combined as four cross-modal stimuli in experiment. 29 subjects participated in the experiment and answered impressions and preference for the combined music piece and scent. Two-way ANOVA showed main effect of scent in one item and main effect of music in three items. Significant Interaction was also observed. These results showed that the main effect of scent exists in the cross-modal stimuli, however, it seems to be weaker than the effect of music pieces.
This paper proposes an indoor positioning method using geomagnetic field with less spatial movement and discusses its positioning performance. In the proposed method, the user holds a geomagnetic sensor, which is often mounted on common smartphones, with his/her arm extended and turns around 360 degrees at his/her current position without walking which is required for traditional methods. The obtained sequence of change in geomagnetic field is used for acquiring the current position. A series of experiments is conducted on positioning accuracy and the results show that the proposed method works with a positioning error of 3.4 m on average and it also demonstrates the best CDF with the positioning error less than about 1.3 m.
Warning alerts are used in situations of the earthquake and so on. In these cases, the warning impression for people is needed to quickly take shelter; however, it sometimes also affords uncomfortableness. The purpose of this study is to propose an Interactive Evolutionary Computation which adjusts the impression of the warning alert by using a set of sound effectors. The target of the search is the appropriate parameters of the effectors on the Earthquake Early Warning alert. Differential Evolution (DE) is used as an evolutionary algorithm, and the efficiency of the Interactive DE for adjusting the impression was investigated through a listening experiment. In the search step, the subjects compared two alerts and selected a better one repeatedly. In the evaluation step, the subjects evaluated the created alerts and the original alert. The results showed a significant increase in fitness by the proposed Interactive DE and effective search processes.
In this study, we propose a system to generate the T-shirt design images that match a user’s preference based on generative adversarial networks (GAN) and a Kansei agent. The generator of GAN can be used to generate lifelike images similar to the learned data. The Kansei agent is a model that learns the user’s Kansei evaluation and retrieves images that the user likes. The Kansei agent is constructed using two neural networks. The first network extracts features from given T-shirt images. The second one estimates the user’s evaluation of T-shirt design using the extracted features. We examined the performance of the proposed system according to the results of the evaluation experiment. The results of the experiment showed that the proposed system finally presented most users T-shirt design images that were highly evaluated.