Curriculum Vitaes

Hiroyuki Kumazawa

  (熊澤 宏之)

Profile Information

Affiliation
Professor, Faculty of System Engineering, Department of System Engineering, Osaka Sangyo University
Degree
Ph. D(Osaka University)
博士(工学)(大阪大学)

Researcher number
00592320
J-GLOBAL ID
201201002457971962
researchmap Member ID
7000000890

Papers

 11
  • Hiroyuki Kumazawa
    11th IEEE International Conference on Sustainable Technology and Engineering, CSDE-2024_606, Dec, 2024  Peer-reviewedLead author
    Drones having high resolution cameras and sensors have become more available and cheaper. We use drone-captured images taken directly above the plantation to research how well the types and areas of fruit trees, as well as other features, can be recognized. Wide-area images are synthesized from numerous locally captured images taken by the drone and are then divided into blocks (image blocks) with a certain amount of overlap to increase the number of blocks available for training. In this paper, semantic segmentation is applied to these blocks, and their classification performance is evaluated. In these evaluations, it is clarified that the overlaps in the blocks make it difficult to properly separate the training data for training the semantic segmentation network from the test data for performance evaluation. To address these issues, data augmentation is applied to the test data, and the evaluation results are presented.
  • 濱田 悠司, 澤 良次, 後藤 幸夫, 熊澤 宏之
    電気学会論文誌C, 130(3) 468-475, Mar, 2010  Peer-reviewed
    In ad-hoc network such as inter-vehicle communication (IVC) system, safety applications that vehicles broadcast the information such as car velocity, position and so on periodically are considered. In these applications, if there are many vehicles broadcast data in a communication area, congestion incurs a problem decreasing communication reliability. We propose autonomous distributed transmit power control method to keep high communication reliability. In this method, each vehicle controls its transmit power using feed back control. Furthermore, we design a communication protocol to realize the proposed method, and we evaluate the effectiveness of proposed method using computer simulation.
  • 伊川雅彦, 五十嵐雄治, 後藤幸夫, 熊澤宏之, 津田喜秋, 森田茂樹
    情報処理学会論文誌, 50(1) 42-50, Jan, 2009  Peer-reviewed
  • 西馬功泰, 後藤幸夫, 熊澤宏之, 駒谷喜代俊
    計測自動制御学会論文集829—836, 42(7) 829-836, Jul, 2006  Peer-reviewed
    We propose a general prediction method based on the efficient computation and online update of the Singular Value Decomposition (SVD) of historical data. The SVD is fundamental to many data modeling algorithms, but the traditional methods for computing it require large computational costs. By adopting a fast sequential SVD updating scheme, the tasks of prediction, imputation of missing values, and model updating can be performed very quickly. In this paper, an application of our method to route travel time prediction is described. Using real travel time data from short sections (links) on expressway, we evaluated prediction performance of travel time on longer section (route) including the links. Experimental comparisons with several statistical machine learning methods suggest that our linear prediction method can achieve similar prediction performance (prediction error) to other nonlinear methods at less computaional cost.
  • 伊川雅彦, 後藤幸夫, 熊澤宏之, 津田喜秋, 岡賢一郎
    電子情報通信学会論文誌, J88-A(2) 218-227, Feb, 2005  Peer-reviewed
  • D Nikovski, N Nishiuma, Y Goto, H Kumazawa
    2005 IEEE INTELLIGENT TRANSPORTATION SYSTEMS CONFERENCE (ITSC), 1074-1079, 2005  Peer-reviewed
    This paper presents an experimental comparison of several statistical machine learning methods for short-term prediction of travel times on road segments. The comparison includes linear regression, neural networks, regression trees, k-nearest neighbors, and locally-weighted regression, tested on the same historical data. In spite of the expected superiority of non-linear methods over Unear regression, the only non-linear method that could consistently outperform linear regression was locally-weighted regression. This suggests that novel iterative linear regression algorithms should be a preferred prediction methods for large-scale travel time prediction.
  • 伊川雅彦, 後藤幸夫, 熊澤宏之, 古澤春樹
    情報処理学会論文誌 2805—2814, 45(12) 2805-2814, Dec, 2004  Peer-reviewed
  • M Ikawa, Y Goto, Y Igarashi, H Kumazawa, K Koizumi, K Oka
    2004 IEEE INTELLIGENT VEHICLES SYMPOSIUM, 105-110, 2004  Peer-reviewed
    In this paper, we show the architecture of DSRC (Dedicated Short Range Communication system) local communication and its application to the information push service. At first, we describe the design goal, architecture and performances of Non-IP part of Application Sub-Layer(ASL), which is the multiple application platform on DSRC in Japan, as DSRC local communication platform. Next, we propose the architecture and the application protocol of the information push service on DSRC as one of the key applications of DSRC. Lastly we report its performance measured through experiments.
  • 熊澤宏之, 富田悟
    電子情報通信学会論文誌1670-1677, J74-D-II(12) 1670-1677, Dec, 1991  Peer-reviewed
  • 熊澤宏之, 富田悟, 中田正矩
    電子情報通信学会論文誌(D-II), J72-D-II(5) 715-723, May, 1989  Peer-reviewed
  • 熊澤宏之, 笠原正雄, 滑川敏彦
    電子情報通信学会論文誌(B), J67-B(1) 1-8, Jan, 1984  Peer-reviewed

Misc.

 3

Presentations

 41
  • Yuuki Fujimoto, Hiroyuki Kumazawa (Osaka Sangyo University), Aki Endo (Mukogawa Women's University)
    The Technical Meeting on "Information Systems", IEE Japan, Feb 25, 2025, The Institute of Electrical Engineers of Japan (IEEJ)
    This study aims to improve the detection accuracy of human body skeletons from crowd images using YOLOv8-pose. The image data used as crowd images are from educational settings where students are performing group exercises such as gymnastics. The research focuses on the image size processed by YOLOv8-pose and examines the differences in detection accuracy based on varying image sizes. The results indicate that for individuals who are close to the camera and occupy a large area, processing with a resized image is preferable, while individuals who are farther away and occupy a smaller area should not have their images resized. By combining these approaches, the goal is to enhance detection accuracy.
  • Hiroyuki Kumazawa
    The 2024 Annual Conference on Electronics, Information and Systems, The Institute of Electrical Engineers of Japan (IEEJ), OS1-2-1, Sep 5, 2024, The Institute of Electrical Engineers of Japan (IEEJ)
    We use the images taken by a drone from directly above a plantation to research on how well the types and areas of the fruit trees and other features can be recognized. Wide area images are synthesized from many locally captured images from the drone and divided into blocks with a certain amount of overlaps to increase the number of blocks for training. In this paper, semantic segmentation is applied to these blocks and their classification performances are evaluated. In these evaluations, it is clarified that the overlaps of the blocks make it difficult to separate the training data to train the network of semantic segmentation and the test data for the performance evaluation. To resolve the problems, data augmentation is applied to the test data and their evaluation results are shown.
  • Hiroyuki Kumazawa
    2023 Annual Conference on Electronics, Information and Systems, Institute of Electrical Engineers of Japan, Aug 30, 2023, Institute of Electrical Engineers of Japan
    Drones having high resolution cameras and sensors have become more available and cheaper. This makes it possible to use them to take images at close range from directly above, which would normally be very hard to take. We used the images taken by a drone from directly above a plantation to research on how well the types and area the fruit trees and other features occupy can be recognized. We reported the application of the deep learning, specifically semantic segmentation, for the monitor of the land usage, which has high accuracy and outstanding results in recognizing objects in the images. In this paper, classification accuracies are evaluated for the various parameters of image blocks, which are extracted from wide area images and used as training data set for the semantic segmentation.
  • Hiroyuki Kumazawa
    IEEJ Annual Conference on Electronics Information and Systems (IEEJ EIS 2022), Aug 31, 2022
    Drones having high resolution cameras and sensors have become more available and cheaper. This makes it possible to use them to take images at close range from directly above, which would normally be very hard to take. We used the images taken by a drone from directly above a plantation to research on how well the types and area the fruit trees and other features occupy can be recognized. In this paper, we apply the deep learning, specifically semantic segmentation, for the monitor of the land usage, which is pixel-by-pixel analysis of an image and decides what class each pixel belongs to. That means the fine classification is possible. In this paper, we show how to capture images and what can be realized using such images. Then classification results of the types and area the fruit trees and other features are shown by using semantic segmentation with various image block sizes.
  • Hiroyuki Kumazawa
    2021 Annual Conference on Electronics, Information and Systems, Institute of Electrical Engineers of Japan, Sep 15, 2021, The Institute of Electrical Engineers of Japan
    We have been considering methods to detect transportation modes using sensor data such as GPS, acceleration, and gyroscope data collected from smartphones. Transportation modes include walk, vehicle, bus, train, and bike. In the previous papers, we presented transportation mode detection from acceleration and gyroscope data using machine learning such as decision tree and random forest, and the method for correcting the classification errors by the post processing of the results of machine learning, which leads to the improvement of the accuracy of the detection. In this paper, we consider the application of LSTM (Long Short Term Memory), which can deal with time series data, aiming at replacing the post processing. At this moment, the LSTM cannot attain the performance of the post processing due to many factors such as the setting of many parameters of LSTM, feature values, and so on, which will be the future works.

Research Projects

 2

研究テーマ

 1
  • 研究テーマ(英語)
    プローブ情報システム
    キーワード(英語)
    プローブ情報、高度道路交通システム、インターネット、通信プロトコル、センシング
    概要(英語)
    スマートフォン、タブレットなど種々センサと通信機能を搭載したデバイスの普及が進んでいる。これらのデバイスでは、センサと通信機能を結びつけることで、自ら移動するセンサとしての機能を有することになる。その機能を活用して様々な情報をセンタに集約するシステムを構築する際の課題、集約した情報の蓄積・解析に関わる課題についての研究を行っている。
    研究期間(開始)(英語)
    2013/04/01