ABSTRACT. The keypoint patches (KPP) extraction is a kind of downsampling method that extracts only feature points and their neighborhoods in a givin point cloud. While it can reduce the computational time for registration, it has a problem that the accuracy of registration deteriorates if keypoints are located in the non-overlapping positions among the point clouds. To solve this problem, a method to dynamically select suitable KPPs from the amount of pre-extracted KPPs using a Genetic Algorithm is proposed. In the evaluation, the selected KPPs are tentative registered using ICP and the transformation that overlaps the most KPPs is evaluated. In this paper, we propose the registration method which uses the suitable initial position of ICP for evaluation. Through the experiments comparing tree methods to determine the initial positions, we verifies that the proposed method enables fast and accurate registration even the overlapping rate is low between point clouds.

ABSTRACT. Planes are simple shapes often found in urban environments. Their detection is crucial for several applications of 3D point clouds. Efforts have focused on detecting planes in low-cost and low-range scenarios. However, in Terrestrial Laser Scanning (TLS), the range is extended to a few hundred meters. This long-range adds huge variations in points' density that complicates defining proper thresholds for conventional methods. Therefore, we propose to employ the sliding voxel plane detector over multiple voxel sizes to estimate hypothetical planes in a coarse-to-fine way, followed by a merging Non-Maximum Suppression. The proposed method shows superior accuracy and efficiency in realistic TLS point clouds.

ABSTRACT. Now, video viewing opportunities are increasing in various situations such as when staying at home and when going outside. With the expansion of 5G services, these op- portunities are expected to increase in the future. Therefore, since it is important to control the amount of transmitted video data, cached video content management methods us- ing priority estimation for video delivery systems with multi- qualities have been studied in one cache server. Since we have focused only on the number of accesses so far, we have not considered the user behavior sufficiently. The prediction of future usage is insufficient, and there is still room for im- provement in cache efficiency. Then, in this study, we pro- pose a cached video content management method in multiple cache servers. In particular, we focus on user behavior and preference, i.e., the placements of cache servers and the us- agevideobias(e.g., videocategoriesandseriesofvideos)are used. Whereas the conventional video management method uses the priority calculations based on an LFU (Least Fre- quency Used) method in each cache, the proposed method does new calculations based on a modified LFU estimation for user behavior and usage video weighting for user prefer- ence in multiple caches. We also design a video usage model and an video access model based on simple daily user be- havior and preference. By the simulation experiments, we evaluate the proposed method by using the cache hit rate for each time interval. By the results, it was revealed that the proposed method could enhance the efficiency by 18.52% compared with the conventional method. We could found the performance improvement of the proposed method.

ABSTRACT. Motion blur is one of the most common causes of image degradation. When restoring an image degraded by motion blur, it is difficult to estimate a PSF and its ideal image from a single input image because the PSF is unknown. To solve this problem, a method that alternates between the PSF estimation and the ideal image estimation has been successfully applied. However, as the resolution of imaging devices improves, it is necessary to deal with larger motion blur. In addition, the circulation of action cameras has increased the conventional method of iterative estimation cannot cope with complex motion blur caused by the motion of the subject. Therefore, in this paper we propose a method based on Tao et al. method to deal with severe motion blur that cannot be recovered by the conventional iterative estimation method and learning-based methods.

ABSTRACT. We have been building a keyword-based virtual museum that allows users to search a database of artworks by giving search keywords, pick up the artworks, and automatically display them in a CG museum space of variable dimension. This time, we have added an annotation system to this museum, allowing users to freely annotate the artworks while enjoying the exhibition. In this way, the database of keyword search is naturally updated, and the descriptions of the artworks freely viewed in the virtual exhibition is elaborated over time. In this way, while enjoying the exhibition, the museum itself becomes richer and evolves on its own, which we call the Growing Museum.

ABSTRACT. Chibanian, the geomagnetism reversal stratum facing the Yorogawa River in Ichihara City, Chiba Prefecture has been announced as the geological age name of the Cenozoic Pleistocene Pleistocene. Then some ideas of local geopark concept have been discussed for the regional activation, and discussion of political civil engineering construction tends to be preceded. In view of another perspective, virtual museum, which will be suitable for a regional activity as inexpensive volunteer effort may be considered. In this paper, constituent requirements for the virtual museum, and some possible technologies have been discussed.

ABSTRACT. This paper discusses the cultural heritages stored in the Central Sulawesi Museum in Palu, Indonesia. Typically, these collections are displayed in the collection room with relatively limited information. Museum visitors generally want to get more information regarding the origin, materials, or other aspects considering that collection. If the cultural heritage is systematically arranged, it is ensured that it will provide more exciting information. This research utilizes a deep learning framework to explore new ways to represent the museum collections. We focus on the classification of cultural heritages in the public Museum in Central Sulawesi, Indonesia. There are about 7.595 item collections classify into ten categories. In this research, we observed 8 of 10 categories only due to the very limited available data to be acquired. We confirmed that our approach shows adequate performance along with all the categories in sense of very few data to be trained. \vspace{1mm}

ABSTRACT. A finger motion tracking display system is developed for Japanese sign language demonstration. The system uses a finger tracking device by infrared stereo cameras (LEAP Motion). This device and its software detects finger shapes and positions in real time. In a current demonstration system, animation pictures are displayed depending on user's finger gestures as Japanese sign language although the number of words are few. The system can also be used for a user interface of interactive museum exibitions without touching.

ABSTRACT. Recently, a lot of live performances come to be distributed on line via the Internet due to the limitation of real live performances on site in COVID-19 situation. Such online performances, however, seem to provide lower impression than the real ones because of the lack of participation feelings in audiences. To make improvements of the participation feelings in online live performances, this article focuses on communication between a performer and audiences and proposes a concept of a new style live distribution with interaction for online live performances based on a result of basic communication experiments with camera.

ABSTRACT. We have prepared organic conducting nanocomposite inks which consist of polypyrrole and indium tin oxide (ITO). The nanocomposite inks of polypyrrole and ITO represent potentially useful processable forms of polypyrrole, normally intractable conducting polymers. The conductivity of the polypyrrole-ITO nanocomposite inks was the same as that of bulk polypyrrole. We confirmed that the polypyrrole-ITO nanocomposites have potential to be utilized as inks for printed electronics due to their high conductivities and colloid stabilities.

ABSTRACT. The distribution of easily imitated counterfeit products, such as food packages, brand　name tags, and pharmaceutical labels, has become a serious economic and safety problem. To address this problem, we propose　a system that can perform the authenticity judgment of genuine and counterfeit products at high speed and accuracy, focusing on the physically unclonable function of inkjet-printed code and locally likely arrangement hashing (LLAH) system, which performs high-speed image retrieval. In this study, we verify the system has high discrimination and stability by obtaining highly accurate results from dataset of up to 4,000 sheets.

ABSTRACT. Many tourists take the city bus to visit national treasure sites in Kyoto. In this study, we propose an application that can manage tourist spot and bus service information. With our application, it is possible to access the same information offered by bus companies by simply touching the bus stop on the street map or on the bus route map displayed on the phone. In addition, GPS displays the user’s present location on the map, which makes it easier to tell where the nearest bus stop is. It lets them know if there is a bus route that goes directly to their destinations or if the desired bus is coming shortly. It also offers easily-browsable basic information about their destinations before reaching them. To make this technically possible, we use Android to create street-map- or bus-route-map pages and implement functions for necessary information to pop up and for page transition, which can be activated by touch operation. The data itself is linked to existing websites, and their key words have been extracted in advance. We define such data as general-purpose data structure. The prototype application has received a lot of favorable feedback.

ABSTRACT. We propose a new e-learning application to promotedigital-transformative education in classical literature classes. To encourage the studies of Hyakunin Isshu, a famous Japanese collection ofWaka poems, this application uses a quest-game format. The basic ideaof the game is that a player tries to navigate a virtual labyrinth full of Hyakunin Isshu picture cards by answering related questions. In the process, the player is expected to learn about the background of the poems and the writers. This application, where Android’s Bitmap drawing function is used to display a two-dimensional labyrinth three-dimensionally, identifies the location of the player within the labyrinth and manages questions related to randomly-placed picture cards as well as the player’s performance at answering them. One of its key technical features is the function to convert the shapes of the picture card images from rectangles to trapeziums (or trapezoids in American English) following the perspective technique. We have also created a PC version of this application where JavaScript is used for implementation to achieve an equivalent effect. Their prototypes are generating a lot of interest among their users.

ABSTRACT. In this research, the researcher proposes a visualized online career matrix that describes university students’ life from birth to their most recent year every three years and describes the interrelationships of each division in a matrix format. In this research, based on the theory of career anchors, which states that it is necessary for career education to deepen self-understanding in the long and short term, in addition to the conventional method of analyzing life stories, a method called career matrix is proposed and a model program is introduced. Then, the effect through the transitions of students’ self-understanding is considered.

ABSTRACT. In Germany, unless people pass the compulsory education graduation exam, they can not go on to the next vocational school. In other words, they can not work independently and become independent workers. It is the purpose of this summer academy to utilize and teach instruction early. Therefore　the target is the 8th grade student (junior high school student in Japan) who is a low-skilled young person without basic academic ability. It supports them to pass the graduation test of compulsory education, go on to vocational school and work independently in society as a professional person finally. Approximately 97% of the participants passed the compulsory education graduation exam, many of them have increased their self-efficacy, and have advanced to vocational schools. The results of investigation on its contents and success factors are reported.