| 立花 幸子 | 1551059: M, 1回目発表 | 知能システム制御 |
|
title: *** Tracking Control of Mechanical Systems using Prediction Governors *** abstract: *** This paper focuses on the tracking control of mechanical systems under the case that the reference signal is unknown. In such case, since a predicted reference signal is used, its prediction error degrades the control performance of the systems. To overcome this problem, we adopt an optimal prediction governor which shapes predicted signals adequately. In this paper, the effectiveness of the proposed method is verified by an experiment with a ball and beam system, which is one of the mechanical systems. *** language of the presentation: *** Japanese (choose one) *** | ||
| 児山 昂生 | 1551045: M, 1回目発表 | 生体医用画像 |
|
title: Analysis of musculoskeletal dynamics around the hip joint from x-ray images on a large-scale cohort database
abstract: In orthopedic surgery, patient-specific musculoskeletal information is essential for diagnosis and therapy. Conventionally, analysis using 2D x-ray images have been performed. However, it is difficult to grasp three-dimensional form and structure. In contrast, three-dimensional analysis can grasp the target from various angles. Three-dimensional analysis of patient-specific musculoskeletal from 2D x-ray images can help surgical planning and rehabilitation. The purpose of this study is to automatically estimate hip posture using 2D-3D registration and analyze the movement of each skeletal muscle from the x-ray images at various postures and a CT image. In addition, we will perform evaluation of our automatic registration algorithm on a large-scale cohort database. In this presentation, we show the result of simulation experiment to verify the robustness of 2D-3D registration algorithm. language of the presentation: Japanese 発表題目: 大規模患者データベースにおける単純X線画像からの股関節筋骨格の動態解析 発表概要: 整形外科領域において、患者固有の筋骨格形態やその動態の解析は診断や治療に不可欠である。従来は、X線画像による解析が行われてきたが、2次元画像による解析では一方向からの情報しか得られず、立体的な形態の把握は困難である。これに対して3次元的な解析では様々な角度からの対象部位の解析が可能となる。そこで臨床に広く用いられている単純X線画像を用いて、3次元的な患者固有の筋骨格の解析ができれば、最適な治療やリハビリテーションの立案に役立つと考える。本研究では、2D-3Dレジストレーションを用いて、単純X線画像とCT画像から股関節の肢位を自動で推定し、複数の肢位での3次元的な筋肉の形状推定および筋動態解析を目的とする。加えて、大規模データを対象に統計的解析を行うためにレジストレーションアルゴリズムの実証実験を行う。本発表では、その初期的検討として、骨盤のCTデータを用いて、2D-3Dレジストレーションの精度およびロバスト性を検証するためのシミュレーション実験を行った結果を示す。 | ||
| 洲崎 智人 | 1551054: M, 1回目発表 | 数理情報学 |
|
title: Modeling of steering control in curve driving
abstract: Drivers’ eye movements during curve is focused in a small region in the visual field.But it is still unknown where drivers' gaze fall in during curve driving and how they decide driving behavior from the visual information. Our study aims to unveil the point where driver direct their gaze and to model the control strategy of drivers in curve driving. In this presentation, I will introduce the related works and our future direction. language of the presentation: Japanese | ||
| 馬場 隆造 | 1551082: M, 1回目発表 | ロボティクス |
|
title:Acquisition of semantic objects and their physical properties by motion analysis
abstract:To enable robots to support humans and use objects, we need to evaluate the physical properties of the object to operate, such as its size and weight, as well as its method of operation. In this paper, we propose a method to automatically acquire a semantic object with its physical properties through object recognition by measurement and analysis of human motion. Finally, we use the semantic object to operate a robot with the actual object. language of the presentation:Japanese 発表題目:動作解析による物性を含んだセマンティックオブジェクトの獲得 発表概要:ロボットが人間と同じ環境内で人間のサポートを行い，様々な物体を使えるようにするため，操作物体の大きさや重さといった物性情報とその物体の操作方法を事前に教える必要がある．本研究では人の動作計測と解析を用いて操作物体を認識，その物性情報を含んだセマンティックオブジェクトを自動的に獲得する手法を提案し，ロボットによる物体の操作に利用する． | ||
| 湯口 彰重 | 1551116: M, 1回目発表 | ロボティクス |
|
title: Unconscious body gestures for an autonomous Android robot
abstract: Android robots with human-like apperance are widely used as communication robots. The motions of Android robots are required to be human-like as they use body gestures to better communicate with humans. On conventional autonomous Android robots, body gestures with explict meaning (e.g., pointing at something) have been proposed but unconscious body gestures (e.g., crossing the arms) have not been deeply studied yet. In this work, we aim to achieve unconscious body gestures on an autonomous Android robot. language of the presentation: Japanese 発表題目: 自律型アンドロイドロボットのための無意識のジェスチャ動作の実現 発表概要: アンドロイドロボットは，外見が人間に酷似しているという特徴を生かし，コミュニケーションロボットとして広く利用されている．コミュニケーションの一つであるジェスチャ動作をアンドロイドロボットで実現する際には，人間に似た動作が求められる．従来の自律型アンドロイドでは，お辞儀など明確な意図を持つジェスチャ動作は実装されてきたが，腕組みなど無意識におこなわれるジェスチャ動作はあまり考慮されてこなかった．本研究では無意識のジェスチャ動作を挿入することでより人間的な動作に近づくという仮説のもと，無意識のジェスチャ動作の実現を目指す． | ||
| 上村 康輔 | 1551035: M, 1回目発表 | ソフトウェア工学 |
|
title: Plagiarism Detection in Binary Programs Using Program Compression
abstract: Open source software (OSS) is obtainable and usable for anyone. On the other hand, its source code can be easily plagiarized. If a developer finds commercial software that seems to contain plagiarized code, it is very difficult to proof the evidence of plagiarism because software is in binary code. Our goal is to detect plagiarism between given two binary programs. We employ a program compression approach, which compares the size between (1) concatenated programs after each program's compression and (2) a compressed program after concatenation of two programs. The method has a potential to very easily detect plagiarism without any program analysis technique. In this study, by using many OSS products, we are going extend the conventional approach by splitting programs into many pieces before compression, and determine the threshold to distinguish between plagiarized program pairs and non-plagiarized pairs. | ||
| 宇野 真純 | 1551018: M, 1回目発表 | 情報基盤システム学 |
|
title:Detection of Drive-by-Download attacks
abstract:In the few years, Drive-by Download attacks is tending to increase, and it's recognized as social problem. It's important to detect a Drive-by Download attack to stop expansion of damage. After introducing the preceding research about drive-by-download attack, I'll report progress and plan of this study. language of the presentation: Japanese | ||
| 秦 恭史 | 1551080: M, 1回目発表 | ユビキタスコンピューティングシステム |
|
title: How is the false and corrected rumors diffused at disaster
abstract: It has been revealed Twitter as a information distribution of the tool is very useful by related research. However, because of its high information diffusion ability, when the rumor spreads, it is necessary to correct them effectively. In this study, we reveal the spread state of false and corrected rumors to diffusion of the information efficiently and correctly. language of the presentation: Japanese 発表題目: 災害時にデマとデマ訂正はどのように拡散するのか 発表概要: 災害時、情報流通のツールとしてTwitterは非常に有用であることが関連研究により明らかになっている。しかし、その情報拡散能力の高さゆえにデマが広がった際、それを効果的に訂正しなければならない。本研究では、効率良く、正しく情報の拡散を行うためにデマとデマ訂正の拡散状況を明らかにする。 | ||
| 小松 巡 | 1451201: M, 1回目発表 | ディペンダブルシステム学 |
| 榎本 陽一 | 1551019: M, 1回目発表 | 自然言語処理学 |
|
title:Parsing of Abstract Meaning Representation
abstract: Abstract Meaning Representation(AMR) is English sentences paired with semantic representation encoded by graphs.The purpose of this research is to make parser that predict AMR graph for given sentence. In this seminar presentation,I will introduce previous researches and my future plan of my research. language of the presentation:Japanese | ||
| 平岡 類 | 1551086: M, 1回目発表 | 知能コミュニケーション |
|
title: Detection of problematic words measuring gaze behavior
abstract: Reading in foreign language is often problematic and time costing , particularly ,when the text contains technical terms . The focus of this research is estimating user's ability and detecting problematic words measuring human gaze behaivor while reading texts. Additionally, utilizing the function of detecting problematic words, we are going to implement a function that displays simplified or translated words in response to input words. language of the presentation: Japanese | ||
| 南村 敏弥 | 1551070: M, 1回目発表 | 視覚情報メディア |
|
title: Image Completion using Planar Structure Guidance
abstract: We propose a method for automatically guiding patch-based image completion using mid-level structural cues. Our method first estimates planar projection parameters, softly segments the known region into planes, and discovers translational regularity within these planes. This information is then converted into soft constraints for the low-level completion algorithm by defining prior probabilities for patch offsets and transformations. Our method handles multiple planes, and in the absence of any detected planes falls back to a baseline fronto-parallel image completion algorithm. We validate our technique through extensive comparisons with state-of-the-art algorithms on a variety of scenes. language of the presentation: Japanese | ||
| 永井 優城 | 1551065: M, 1回目発表 | 自然言語処理学 |
|
title:The research of Text Simplification
abstract: I will talk about the research of Text Simplification. language of the presentation: Japanese | ||
| 國吉 房貴 | 1551040: M, 1回目発表 | 光メディアインタフェース |
|
title: Refocus Image Generation for Lens-less Microscope
abstract: Recently, lens-less microscope receives more attention from industrial usage. However, it is difficult to observe 3d-shape of subject because there is no depth blur in an acquired image using a lens-less microscope. Thus, we propose a method to generate a refocus image from the lens-less microscope. Acquiring the light-field by illuminating from various directional light sources, we generate refocus image from the light-field. Furthermore, we apply our method to real world environment to indicate the effectiveness of our method. language of the presentation: Japanese 発表題目: レンズレス顕微鏡を用いたリフォーカス手法の提案 発表概要: 近年注目を浴びている撮像系からレンズを省いたレンズレス顕微鏡は，被写体の奥行に関わらずボケが生じないため，被写体の立体的な構造が把握しにくいという問題がある． そこで本研究では，異なる方向の光源から照明された被写体を撮像することで得られたシーンの光線空間を用いて，奥行によるボケを有するリフォーカス画像を生成するアルゴリズムを提案する． さらに，実際にレンズレス顕微鏡で撮影された被写体に対してリフォーカス画像を生成することで手法の有効性を実証した． | ||