研究者詳細 - 米澤　拓郎

論文 - 米澤　拓郎

分割表示 301 件中 61 - 80 件目／全件表示 >>

スマートフォン・全方位カメラを用いた作業効率化のための半自動天伏図作成システム査読有り

小濱大輝, 永田吉輝, 安武和成, 酒見和幸, 片山晋, 浦野健太, 米澤拓郎, 河口信夫

空気調和・衛生工学会大会　学術講演論文集 2024.10 巻 ( 0 ) 頁： 485 - 488 2024年

記述言語：日本語掲載種別：研究論文（学術雑誌）出版者・発行元：公益社団法人空気調和・衛生工学会

建物の改修工事の際，現状の図面が必要となる．10年以上運用された建物では，図面がなかったり，あったとしても過去の改修を反映していなかったりする．天井設備が記載された天伏図の修正のために現地調査がよく行われる．しかし，天井設備の位置や種類を正確に確認するのは手間が掛かる．本研究では，天伏図作成の作業効率を高めるために，スマートフォン・全方位カメラで撮影した動画から，既存建物の天伏図を少ない作業時間で半自動的に作成するシステムを提案する．

DOI： 10.18948/shasetaikai.2024.10.0_485

CiNii Research

Performance Evaluation of KNIME Low Code Platform in Deep Learning Study and Optimal Hyperparameter Tuning 査読有り

Thongkhome P., Yonezawa T., Kawaguchi N.

IEEE Region 10 Annual International Conference Proceedings TENCON 頁： 1373 - 1376 2024年

記述言語：日本語掲載種別：研究論文（学術雑誌）出版者・発行元：IEEE Region 10 Annual International Conference Proceedings TENCON

A low-code platform is a software development environment that allows for the creation of applications through graphical user interfaces and configuration instead of traditional hand-coded computer programming. In this study, the application to classify a dataset of traffic sign images using the KNIME low-code deep learning development platform will be discussed to represents this software performance especial in term of model optimization processes. By creates the workflow to perform image preprocessing, create the CNN layer under KERAS sequential API and finding the best set of key hyperparameters among traditional KNIME build-in optimization algorithm including Brute force, Hill climbing, random search, Bayesian Optimization and black-box optimizer Optuna optimization algorithm under 3 types CNN architecture as simple CNN, Resnet-50 and VGG16 to classify traffic sign images. The result demonstrates that both grid search and random search optimization can be effective, while both Optuna and Bayesian optimization stands out as a powerful method due to its ability to efficiently explore the hyperparameter space and achieve superior results to meet 99% accuracy under simple CNN environment, but Optuna is significantly improve optimization times than Bayesian about 7 - 8 times. The KNIME low-code platform provides a user-friendly environment for developing and fine-tuning models to contribute the ongoing progress in machine learning and deep learning research development.

DOI： 10.1109/TENCON61640.2024.10903074

Scopus

Trusted Inter-Reality基盤における利用主体に関するトラスト招待有り査読有り Open Access

金岡晃, 米澤拓郎

情報処理学会論文誌 64 巻 ( 12 ) 頁： 1590 - 1598 2023年10月

CrowdFlowTransformer: Capturing Spatio-Temporal Dependence for Forecasting Human Mobility 招待有り

Choya T, Tamura N, Katayama S, Urano K, Yonezawa T, Kawaguchi N

2023 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2023 頁： 496 - 501 2023年3月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：2023 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, PerCom Workshops 2023

Crowd flow forecasting is expected to have a wide range of applications such as human resource allocation, guidance design, marketing, disaster mitigation and congestion prediction for avoiding epidemic such as COVID-19. Crowd flow forecasting is challenging because it requires considering both the task of capturing the temporal dependency of data and capturing the spatial dependence. To address these challenges, in this paper, we propose a mechanism for referencing time-series features that are important for forecasting and incorporating graph convolution into Transformer, and we introduce CrowdFlowTransformer(CF-Transformer), a deep learning model based on Google's Transformer framework captures the Spatio-temporal dependency of time series. CF-Transformer captures the time series dependency by extracting important local time series from the past time series, inputting them to the decoder of Transformer, and encoding critical features into the model's input. We adapted CF-Transformer to a real-world crowd flow dataset. We evaluated it by comparing its forecasting accuracy with conventional models, and the results demonstrate that our model outperforms the conventional models.

DOI： 10.1109/PerComWorkshops56833.2023.10150301

Scopus

Understanding Regional Characteristics Through EC Data Analysis 査読有り

Yamaguchi, K; Shoji, K; Tamura, N; Sakakura, N; Matsukura, Y; Hiranaga, Y; Shimomura, K; Urano, K; Yonezawa, T; Kawaguchi, N

DISTRIBUTED, AMBIENT AND PERVASIVE INTERACTIONS, DAPI 2023, PT I 14036 巻頁： 390 - 404 2023年

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

The COVID-19 pandemic, which began in 2020, has changed people’s lives, and people are shopping more online. While the analysis of online shopping is becoming increasingly important, regional differences in consumption trends exist. This study proposes a data-driven regional modeling method based on EC purchase data to examine the regional characteristics of online shopping purchase trends. Using the proposed method, we quantified and visualized the degree of similarity of consumption trends among regions and the degree of dispersion of consumption trends within regions using approximately 300,000 lines of online shopping history for 3 years in Japan. As a result, we found that there was some disruption in the early stages of e-commerce for food products by region, while there was little difference in consumption trends among regions for daily necessities. In addition, for consumer durables and clothing, regional differences in consumption trends were confirmed in terms of the number of cars owned per capita, urbanization status, and other regional characteristics.

DOI： 10.1007/978-3-031-34668-2_26

Web of Science

Scopus

Panoramic Ceiling Image Synthesis Method Prioritizing Fixture Outlines using an Omnidirectional Camera 査読有り

Kohama D., Nagata Y., Yasutake K., Urano K., Katayama S., Yonezawa T., Kawaguchi N.

2023 14th International Conference on Mobile Computing and Ubiquitous Network Icmu 2023 頁： 1 - 8 2023年

記述言語：日本語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：2023 14th International Conference on Mobile Computing and Ubiquitous Network Icmu 2023

The manual creation of a ceiling plan consumes many human resources to confirm the current state of existing buildings for the renovation. Especially the positions and the types of existing fixtures are important. In this paper, we propose a synthesis method for a panoramic ceiling image using a video shot by an omnidirectional camera. We utilize a Visual SLAM method to map the positions of omnidirectional video frames including the ceiling. We prioritize the clear depiction of fixture outlines for easier identification of the types of ceiling fixtures. To draw clear fixture outlines in the resulting image, each fixture is referred from the single video frame that captures it most clearly. In the experiment, we collected an omnidirectional video in the indoor area of approximately 273 square meters with multiple rooms. We evaluated the synthesized image for positional accuracy of the ceiling fixtures, and our proposed method demonstrated higher accuracy compared to the baseline.

DOI： 10.23919/ICMU58504.2023.10412256

Scopus

その他リンク： https://dblp.uni-trier.de/db/conf/icmu/icmu2023.html#KohamaNYUKYK23

Panoramic Ceiling Image Synthesis Method Prioritizing Fixture Outlines using an Omnidirectional Camera 査読有り

Kohama, D; Nagata, Y; Yasutake, K; Urano, K; Katayama, S; Yonezawa, T; Kawaguchi, N

2023 FOURTEENTH INTERNATIONAL CONFERENCE ON MOBILE COMPUTING AND UBIQUITOUS NETWORK, ICMU 2023年

Open-Domain Dialogue Management Framework Across Multiple Device for Long-Term Interaction 査読有り

Katayama, S; Hayashida, N; Urano, K; Yonezawa, T; Kawaguchi, N

HUMAN-COMPUTER INTERACTION, HCI 2023, PT III 14013 巻頁： 354 - 365 2023年

記述言語：日本語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Lecture Notes in Computer Science Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics

This study explores the feasibility of dialogue systems with individuality capable of providing continuous and lasting assistance via a multiple device dialogue system. A framework has been devised to manage dialogue history, allowing for the use of a singular identity across various interfaces, including chatbots and virtual avatars. This framework can summarize and save the dialogue history, which can be utilized to generate responses. The impact of dialogue history sharing on users’ interactions with a particular character across various devices was assessed for naturalness, continuity, and reliability. The results indicate that dialogue history sharing can foster more natural and continuous conversations, thereby enhancing the potential for long-term support. This research advances the proposition that a digital agent endowed with a consistent identity across multiple devices can provide personalized and sustained support to users.

DOI： 10.1007/978-3-031-35602-5_25

Web of Science

Scopus

その他リンク： https://dblp.uni-trier.de/db/conf/hci/hci2023-3.html#KatayamaHUYK23

Digitization and Analysis Framework for Warehouse Truck Berth 査読有り

Yokoyama, K; Katayama, S; Urano, K; Yonezawa, T; Kawaguchi, N

2023 FOURTEENTH INTERNATIONAL CONFERENCE ON MOBILE COMPUTING AND UBIQUITOUS NETWORK, ICMU 2023年

Digitization and Analysis Framework for Warehouse Truck Berth 査読有り

Yokoyama K., Katayama S., Urano K., Yonezawa T., Kawaguchi N.

2023 14th International Conference on Mobile Computing and Ubiquitous Network Icmu 2023 2023年

記述言語：日本語掲載種別：研究論文（学術雑誌）出版者・発行元：2023 14th International Conference on Mobile Computing and Ubiquitous Network Icmu 2023

In distribution warehouses, various forms of data transformation are underway, such as the conversion of data related to receiving and inventory management. Distribution warehouses have designated areas known as truck berths, where trucks are parked for loading/unloading. While data collection is expected to enhance work efficiency at these truck berths, there is no data regarding the status of work in progress in the unloading space, and no system is in place to analyze the work. Consequently, it is unclear which tasks are problematic and how to enhance efficiency. To address this issue, this paper proposes a framework for analyzing the loading/unloading situation at truck berths with the objective of enhancing berth utilization efficiency and work efficiency. Specifically, we employed a fixed camera for object detection through instance segmentation. This approach enables us to understand the situation of trucks, unloading spaces, and workers. The training data were annotated using a large amount of real-world data, allowing for the measurement and analysis of the efficiency of loading/unloading operations. By calculating the amount of cargo based on the floor space, we assessed the efficiency of unloading operations per truck. Through the analysis of both efficient and inefficient operations, insights for enhancing work efficiency were provided.

DOI： 10.23919/ICMU58504.2023.10412228

Scopus

Gait-Robust Heading Estimation Using Horizontal Acceleration for Smartphone-based PDR 査読有り

Kano K., Yoshida T., Katayama S., Urano K., Yonezawa T., Kawaguchi N.

Ceur Workshop Proceedings 3581 巻 2023年

記述言語：日本語掲載種別：研究論文（学術雑誌）出版者・発行元：Ceur Workshop Proceedings

This study tackles heading estimation for Pedestrian Dead Reckoning (PDR) with smartphones. In dealing with changes in the holding posture of smartphones, it works to consider the relationship between sensor orientation and heading. However, the existing methods lack robustness to various gaits, such as sideways and backward walking. Therefore, we propose a novel method considering various spatiotemporal features of horizontal acceleration with deep learning. The proposed method calculates horizontal acceleration in the global coordinate system from measured acceleration, gravitational acceleration, and rotation vector. Then, it inputs the horizontal acceleration over a certain period into a deep neural network model and predicts the unit vector directed to the mean heading during that period. We created a dataset covering multiple gaits and evaluated the method using four models: Convolutional Neural Network (CNN), Bidirectional Long Short-Term Memory (BiLSTM), DualCNN-LSTM, and DualCNN-Transformer. Consequently, we found that the proposed method was more robust to gaits than the existing methods, with the DualCNN-LSTM and DualCNN-Transformer models achieving the highest accuracy.

Scopus

Gait-Robust Heading Estimation Using Horizontal Acceleration for Smartphone-based PDR. 査読有り

Kazuma Kano, Takuto Yoshida, Shin Katayama, Kenta Urano, Takuro Yonezawa, Nobuo Kawaguchi

IPIN-WiP 2023年

Mobility Link XR: Interspace Interaction System in Electric Wheelchair 査読有り

Hayashida N., Shimosato H., Urano K., Yonezawa T., Kawaguchi N.

Communications in Computer and Information Science 1833 CCIS 巻頁： 314 - 320 2023年

記述言語：日本語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Communications in Computer and Information Science

This research proposes a communication system called Mobility Link XR that connects physical space and cyberspace with mobility. Mobility Link XR is a system that enables remote users to view panoramic video from a 360-degree camera attached to a mobility vehicle in different space by wearing a VR device, and mobility users to view the remote user as an avatar by wearing an MR device. In this way, sharing space in three dimensions using XR enables a higher level of human communication. In this paper, we apply Mobility Link XR to an electric wheelchair and design two types of scenarios: an assistance mode that reproduces the positional relationship of communication in a conventional wheelchair, and a passenger mode that reproduces the positional relationship inside a vehicle, which is said to be a suitable distance for conversation. We also evaluated the reproducibility of communication in the wheelchair and the effectiveness of communication using the avatar. The results showed that the reproducibility of voice and emotion was highly evaluated and that the side-by-side positional relationship enabled higher quality communication as the avatar was more easily seen and felt present.

DOI： 10.1007/978-3-031-35992-7_43

Scopus

その他リンク： https://dblp.uni-trier.de/db/conf/hci/hci2023-44.html#HayashidaSUYK23

ER-Chat: A Text-to-Text Open-Domain Dialogue Framework for Emotion Regulation 査読有り

Katayama, S; Aoki, S; Yonezawa, T; Okoshi, T; Nakazawa, J; Kawaguchi, N

IEEE TRANSACTIONS ON AFFECTIVE COMPUTING 13 巻 ( 4 ) 頁： 2229 - 2237 2022年10月

記述言語：日本語掲載種別：研究論文（学術雑誌）出版者・発行元：IEEE Transactions on Affective Computing

Emotions are essential for constructing social relationships between humans and interactive systems. Although emotional and empathetic dialogue generation methods have been proposed for dialogue systems, appropriate dialogue involves not only mirroring emotions and always being empathetic but also complex factors such as context. This paper proposes Emotion Regulation Chat (ER-Chat) as an end-to-end dialogue framework for emotion regulation. Emotion regulation is concerned with actions to approach appropriate emotional states. Learning appropriate emotion and intent when responding on the basis of the context of the dialogue enables the generation of more human-like dialogue. We conducted automatic and human evaluations to demonstrate the superiority of ER-Chat over the baseline system. The results show that inclusion of emotion and intent prediction mechanisms enable generation of dialogues with greater fluency, diversity, emotion awareness, and emotion appropriateness, which are greatly preferred by humans.

DOI： 10.1109/TAFFC.2022.3191973

Web of Science

Scopus

テキスト対話システムにおける感情制御を伴う応答生成手法

片山晋, 米澤拓郎, 河口信夫

知能と情報 34 巻 ( 3 ) 頁： 568 - 578 2022年8月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：日本知能情報ファジィ学会

感情を扱う対話システムはユーザの満足度向上や肯定的なインタラクション増加の効果が認められており，人間に寄り添うデジタルパートナーとしての役割が期待されている．対話システムがユーザの感情を認識し，対話システム自身の感情を表現するためには，対話の文脈から適切な感情を選択する感情制御が不可欠である．本研究では，応答時の適切な感情を推定する感情制御を伴うテキスト応答生成フレームワークを提案する．提案手法はユーザの発話テキストにおける意味的文脈と感情的文脈を考慮することで応答時の適切な感情を推定し，感情表現を含んだ応答テキストをニューラルネットワークを用いて生成する．実験では，対話生成の品質に対する評価指標を用いた自動評価と，クラウドソーシングサイトを用いて集めた100人の被験者による人手評価を行い，提案手法の有効性を示した．

DOI： 10.3156/jsoft.34.3_568

CiNii Research

Time-Sensitive Cooperative Perception for Real-Time Data Sharing over Vehicular Communications: Overview, Challenges, and Future Directions 査読有り

Aoki S., Yonezawa T., Kawaguchi N., Steenkiste P., Rajkumar R.R.

IEEE Internet of Things Magazine 5 巻 ( 2 ) 頁： 108 - 113 2022年6月

記述言語：日本語掲載種別：研究論文（学術雑誌）出版者・発行元：IEEE Internet of Things Magazine

Cooperative perception is a prospective application to improve road safety by having connected autonomous vehicles (CAVs) exchange their raw or processed sensor data over vehicular communications. Since CAVs heavily rely on sensor-based perception, including vision cameras, LiDARs, and radars, cooperative perception has an immense potential to improve road safety. At the same time, a variety of sensors and edge servers have been widely deployed in smart cities, and such sensors and servers might be able to empower CAVs on public roads. In this article, we comprehensively study such cooperative perception for overview, technical challenges, practical requirements, prospective system designs, current approaches, and future research directions. In particular, we focus on the time sensitivity of the cooperative perception frameworks, in which delays of computing and communications lead to detection errors. In addition, we comprehensively study sensor fusion frameworks and models to improve the detection accuracy and reliability of CAVs while avoiding information flooding and/or rumor spreading. Since network congestion might lead to packet collisions and/or delays, congestion control might be essential to use cooperative perception in practice. Finally, we discuss the technical and ethical challenges of using cooperative perception on public roads and conclude with future research directions.

DOI： 10.1109/iotm.001.2100143

Scopus

深層学習を用いたIMU付きペンによる手書き文字認識査読有り Open Access

挺屋友幹, 永田吉輝, 東浦圭亮, 下里浩昇, 山口公平, 村井大地, 深谷愓也, 坂倉波輝, 戸出悠太, 浦野健太, 米澤拓郎, 河口信夫

人工知能学会研究会資料知識ベースシステム研究会 125 巻 ( 0 ) 頁： 03 2022年3月

RobotNEST: Toward a Viable Testbed for IoT-Enabled Environments and Connected and Autonomous Robots 査読有り

Aoki, S; Yonezawa, T; Kawaguchi, N

IEEE SENSORS LETTERS 6 巻 ( 2 ) 頁： 1 - 4 2022年2月

記述言語：日本語掲載種別：研究論文（学術雑誌）出版者・発行元：IEEE Sensors Letters

Autonomous mobile robots (AMRs), including delivery robots, security robots, and automated wheelchairs, are very promising for improving the quality of life, safety, mobility, and productivity. To deploy such sensor-rich AMRs in real-world society, we have to resolve not only technical challenges but also ethical and societal issues. In addition, since multiple companies are developing different types of AMRs individually, we have to design and develop practical platform to enable safe cooperation and collaboration before using them in the society. Hence, in this letter, we present a viable Nagoya Experimental Site and Testbed named <italic>RobotNEST</italic> developed at Nagoya City, Japan, where we leverage vehicular communications and edge/cloud computing technologies to develop cooperative mobile robots. RobotNEST provides real-world AMRs, experimental fields, learning kits, and infrastructural sensors and communications, including 5G private networks. We also introduce multiple developed applications at RobotNEST that enable safe cooperation for autonomous robots. Finally, we discuss the scalability and limitations of the testbed and conclude with future research directions.

DOI： 10.1109/LSENS.2021.3139624

Web of Science

Scopus

A Data-Driven Approach for Online Pre-impact Fall Detection with Wearable Devices 査読有り

Yoshida, T; Kano, K; Higashiura, K; Yamaguchi, K; Takigami, K; Urano, K; Aoki, S; Yonezawa, T; Kawaguchi, N

SENSOR- AND VIDEO-BASED ACTIVITY AND BEHAVIOR COMPUTING, ABC 2021 291 巻頁： 133 - 147 2022年

記述言語：日本語掲載種別：論文集(書籍)内論文出版者・発行元：Smart Innovation Systems and Technologies

The implementation of wearable airbags to prevent fall injuries depends on accurate pre-impact fall detection and a clear distinction between activities of daily living (ADL) and them. We propose a novel pre-impact fall detection algorithm that is robust against ambiguous falling activities. We present a data-driven approach to estimate the fall risk from acceleration and angular velocity features and use thresholding techniques to robustly detect a fall before impact. In the experiment, we collect simulated fall data from subjects wearing an inertial sensor on their waist. As a result, we succeeded in significantly improving the accuracy of fall detection from 50.00 to 96.88%, the recall from 18.75 to 93.75%, and the specificity 81.25 to 100.00% over the baseline method.

DOI： 10.1007/978-981-19-0361-8_8

Web of Science

Scopus

MetaPo: A Robotic Meta Portal for Interspace Communication 査読有り

Yonezawa, T; Hayashida, N; Przybilla, J; Kyono, Y; Urano, K; Kawaguchi, N

PROCEEDINGS OF SIGGRAPH 2022 POSTERS, SIGGRAPH 2022 頁： 28 - 2 2022年

記述言語：日本語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Proceedings SIGGRAPH 2022 Posters

We introduce MetaPo, a mobile robot with spheric display, 360° media I/O and robotic hands for creating a unified model of interspace communication. MetaPo works as a portal between pairs of physical-physical, cyber-cyber and cyber-physical spaces to provide 1) panoramic communication for multiple remote users, and 2) immersive interspace migration with mobility functionality. The paper overviews our concept and first prototype of MetaPo with its hardware and software implementation.

DOI： 10.1145/3532719.3543255

Web of Science

Scopus

その他リンク： https://dblp.uni-trier.de/db/conf/siggraph/siggraph2022posters.html#YonezawaHUPKK22

　前のページ - 次のページ　

<米澤 拓郎>

論文 - 米澤 拓郎

<米澤　拓郎>

論文 - 米澤　拓郎