コロキアムB発表

日時: 9月22日(水)5限(16:50~18:20)


会場: L1

司会: 田中 宏季
児玉 哲哉 M, 2回目発表 インタラクティブメディア設計学 加藤 博一, 向川 康博, 神原 誠之, 藤本 雄一郎, 澤邊 太志
title:Accurate 3D pose estimation with monocular camera and accurate 3D human model in 100m sprint scene abstrat:The measurement and analysis of running posture has been conducted to improve the performance of the 100m run in athletics. The conventional method of measuring running posture is to take images or videos. This method is easy to use if you have a camera, but it can only analyze two dimensions from the camera direction. On the other hand, motion capture can measure 3-dimensional posture, so it is possible to perform 3-dimensional analysis from any direction. However, motion capture requires some special equipment and is inconvenient to carry around, so it is not easy to measure. In this research, we aim to establish a technology to accurately estimate the 3D running posture of a 100m track and field athlete using a simple measurement method, assuming that a highly accurate 3D model of the human body and the person's running data have been obtained.
language of the presentation: Japanese
発表題目:陸上100m走における単眼カメラと高精度三次元人体モデルを用いた三次元位置姿勢推定
発表概要:陸上競技の100m走では,パフォーマンス向上のために走行姿勢の計測・分析が行われています.従来の走行姿勢の計測方法として,画像や映像の撮影が挙げられます.この方法はカメラのみで行える手軽な計測方法になりますが,カメラ方向からの二次元的な分析しかできません.これに対して,モーションキャプチャでは,三次元の姿勢を計測できるため,任意方向からの三次元的な分析が行えます.しかし,複数の特殊装置を必要とするため,持ち運びに不便であり手軽な計測が行えません.そこで,本研究では,高精度な人体3Dモデルと個人の走行データが取得してあるという前提条件のもと,手軽な計測方法で陸上100m選手の三次元の走行姿勢を高精度に推定する技術の確立を目指します.
 
吉村 和真 M, 2回目発表 サイバネティクス・リアリティ工学 清川 清, 杉本 謙二, 酒田 信親, 内山 英昭, 磯山 直也
title: Study of a Racket-Type Device that Reproduces Impulsive Force of Ball Games
abstract: Currently, sports experiences using VR are in practical use, but they present only visual information. In order to provide a more realistic experience, we need a device that can reproduce the tactile sensation that occurs when a ball is hit back. However, it is difficult for existing force-feeling devices to present a strong hitting force without restricting the motion. In this study, we will develop a device that can present strong impact force without restricting the motion. In this paper, we report a prototype of a device that can present force without motion limitation.
language of the presentation: Japanese
発表題目: 球技の打撃力を再現するラケット型デバイスの研究
発表概要: 現在,VRを活用したスポーツ体験が実用化されているが,これは視覚情報のみの提示である.よりリアリティのある体験を提供するため,ボールを打ち返した時に生じる触覚の再現が可能なデバイスが必要である.しかし既存の力覚提示デバイスでは,動作の制限がなく強い撃力を提示することは困難である.そこで本研究では動きの制限がされず,強い撃力を提示できるデバイスを開発する.本発表では,動きの制限がなく撃力を提示するデバイスを試作したので報告する.
 
GOMEZ VARGAS ANDRES ROBERTO D, 中間発表 サイバネティクス・リアリティ工学 清川 清, 杉本 謙二, 酒田 信親, 内山 英昭, 磯山 直也
title: High level data based augmented reality assistance for human-robot coexistence
abstract: Human-robot communication in shared spaces is still a largely unexplored topic with many opportunities of improvement. With the increasing use of robots in several work domains, unwanted/unexpected interactions between people and robots are becoming commonplace. Safety and navigation around robots have been identified as the main sources of concern for people in this kind of encounters, making them a priority when considering improving human-robot interactions. With this in mind, augmented robotics have been proposed previously as a medium to allow robots to communicate awareness and motion intent to those around them. We have done a general survey of these proposals, identifying two specific opportunities of improvement and unexplored use cases: Allowing human-robot negotiation of space based on obstacle cost visual representation, and stimulating robot avoidance based on movement implicit guidance. Our final goal is to contribute to the design of a robust toolkit of visual imagery with high level data visual representation.
language of the presentation: English