| 大谷 悠太 | 1551023: M, 1回目発表 | 生体医用画像 |
| title:Initial consideration of building statistical shape model of pelvis for large-scale data including CT data with insufficient FOV. abstract:In building statistical shape models, it is reported that the more training data increase the better performance.So it is important to include a large number of data in building statistical shape models.However there are some lacked edge data because of narrowed field of view in performing the CT scan.Such data can’t be included in training data for statistical shape models, thereby we can’t make full use of the number of the data set.We consider a method for building a statistical shape model from such lacked data by excluding corrupted parts from the model.In this presentation, We will show the result of excluding the lacked pelvis data from statistical shape model. langage of the presentation:Japanese | ||
| 趙 崇貴 | 1551062: M, 1回目発表 | ロボティクス |
|
title: Hand motion recognition by the forearm shape measurement based on distance sensor array Recently, surface electromyogram (sEMG) is used as an input signal in hand motion recognition for interfaces to make intuitive operation of machines possible. However, the information that we can measure with sEMG is only the activity of the muscles of the outer layer part of the skin. It will be necessary to consider not only the outer layer muscles but also the activity of muscles deeper under the skin in the future to recognize many kinds of movements. In this study, I develop a method to recognize movements of the hand by measuring the shape change of the forearms which are a result of the composition movement of both outer and deep layer muscles using a distance sensor array. language of the presentation: Japanese 発表題目 : 距離センサアレイを用いた前腕形状変化計測による手の動作認識 発表概要 : 機械の操作インタフェースへの応用を目的として、表面筋電位を用いた手の動作認識が研究されてきた。しかし、表面筋電位で取得可能な情報は表層筋の活動のみである。さらに複数の手の動作を高精度で認識するためには、表層筋だけではなく深層筋の活動も考慮する必要がある。本研究では、距離センサアレイを用いて前腕の形状変化を計測することによって、表層筋と深層筋の活動に基づく手の動作認識を行う。
| ||
| 岡田 裕斗 | 1551025: M, 1回目発表 | 数理情報学 |
|
Title: The inference of deep brain activity using resting state fMRI and fNIRS
Abstract: Functional magnetic resonance imaging (fMRI) and functional near infrared spectroscopy (fNIRS) are methods for non-invasively measuring human brain activity. Functional connectivity during resting state is used to study functional brain networks. However, fMRI has a disadvantage such as the restriction of subject's action. On the other hand, fNIRS has a disadvantage such as the inability to measure deep brain activity. Therefore, it's required to enable to measure whole brain activity in a natural state. In this study, I verify whether it's possible to estimate deep brain activity from brain cortical activity by combining resting state fMRI and fNIRS. Language of the presentation: Japanese | ||
| 田口 智之 | 1551057: M, 1回目発表 | ロボティクス |
|
title: Detection of occluded regions by the stereo camera for a mobile robot
abstract: Environmental recognition is important to realize the smooth movement of the robot. In this work, environmental recognition consists of recognizing walking paths and walls, and understanding which regions robots and humans can move in. Additionally, in environments where humans and robots are coexisting, I propose a stereo vision system for the purpose of smooth movement, that estimates walking paths in occluded regions based on walls and shadows. In this work, I examine stereo matching techniques for estimating a road surface, detecting the boundary line of walls and things and extracting occluded regions. language of the presentation: Japanese | ||
| 山本 賢佑 | 1551114: M, 1回目発表 | 大規模システム管理 |
|
title: Replication scheduling for the Hadoop distributed file system
abstract:Hadoop, a distributed computing environment framework, makes use of a distributed file system, called HDFS, to store data that one wants to analyze. HDFS divides the data into multiple blocks and replicates the specified number of blocks onto discrete nodes. The replication scheduling significantly affects the efficiency of the distributed computation. In this talk, I will introduce existing researches for replication scheduling when some nodes break down and present the perspective of my research. language of the presentation: Japanese | ||
| 古川 智也 | 1551095: M, 1回目発表 | 情報基盤システム学 |
|
title: Post-Quantum cryptography for embedded systems
abstract: In preparation for the realization of quantum computers, Post-Quantum Cryptography is studied.Post-Quantum Cryptography is cryptography is secure against quantum computers. However, It is different to use in enviroment restricted computer power, capacity memory, and so on like embedded systems. In this presentation, I will explain an overview of Ring-LWE is one of Post-Quantum Cryptography and introduce previous study of Ring-LWE for embedded system. language of the presentation: Japanese 発表題目: 組込みシステムにむけたポスト量子暗号 発表概要: 量子コンピュータの実現に備え，ポスト量子暗号が研究されている．ポスト量子暗号とは量子コンピュータにも安全な暗号のことを示す． しかし，組込みシステムのようなマシンパワー，メモリ容量が制限されている環境で利用するのは難しい．本発表ではポスト量子暗号の一つであるRing-LWEの概要を説明し,組込みシステムに向けたRing-LWEの先行研究を紹介する． | ||
| 山崎 雅也 | 1551112: M, 1回目発表 | ソフトウェア設計学 |
|
title: Orthogonal Defect Classification for Spacecraft Software and its application for Verification Strategy
abstract: It is effective to use anomaly data for that of causing event analysis and improving future software development process. However, there aren’t so many companies use the data as things are stand because they don’t know how to utilize it after classification. In this research, I applied Orthogonal Defect Classification (ODC) that can analyze such as defect zone and mixed process for spacecraft software to find out anomaly occurrence process and the controversial. By using that result, I will aim to construct software verification strategy for developing safer system. In this presentation, I will introduce about ODC, a related work in NASA operation and result of applying ODC to Japan’s spacecraft software. Finally, I will show the direction of my research plan. language of the presentation: Japanese 発表題目: 宇宙機ソフトウェアへの直行欠陥分類法の適用と検証戦略への応用 発表概要: ソフトウェア開発における不具合情報を活用することは、その原因事象の分析や今後の開発工程の改善に向けて有効である。しかしながら、現状では不具合情報は、効率的な分析が困難であることから活用している企業は少ない。その理由として欠陥を分類した後の用途を明確にできていない点にある。本研究では、非常に高い信頼性が求められる宇宙機ソフトウェアに対して、効率的に欠陥領域、混入工程などを分析できるとされる直交欠陥分類法を不具合情報に適用し、ソフトウェアの不具合が発生する過程、問題点を明らかにするとともに、より安全なシステムを開発するためにソフトウェア検証戦略への応用を目指す。本発表では直交欠陥分類法の概要及びNASAでの関連研究を紹介するとともに、実際に日本の宇宙機ソフトウェア開発で発生した不具合情報に直交欠陥分類法を適用した結果を示し、最後に今後の研究方針を述べる。 | ||
| 則行 祐作 | 1551077: M, 1回目発表 | ソフトウェア工学 |
|
title: A Study of the Growth of Programmers by Mining Programming Learning Website Datasets
abstract: An online judge is an online system to test programs in competitive programming. The system can execute code, return the result, and store the histories of executions. Users can learn and improve their algorithmic programming skills by addressing programs in an online judge system. With the dataset of AOJ online judge, we plan to study the growth of users, that is, programmers, by analyzing the characteristics of programs and the histories of the submission results. To analyze the characteristics of programs, we apply item response theory, which is used for designing, analyzing, and scoring of tests. | ||
| 前田 直樹 | 1551099: 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: Proposal of Communication Supporting System Based on Face Recognition Camera abstract: Factor that triggers the take the first meeting of people and communication is needed. Although there is a commemorative photograph as one of the factors, it is difficult to apply for a commemorative photo in the first meeting of the opponent. The system using a face recognition camera make the communication support by commemorative photograph. The system performs the communication induction, and the communication facilitation. System performs communications facilitation by bringing up a subject. And system performs communication induction by demanding commemorative photograph with another human. language of the presentation : Japanese 発表題目: 顔認識カメラを備えたコミュニケーション支援システムの提案 発表概要: 初対面の人物とコミュニケーションを取るには切っ掛けとなる要因が必要である。要因の一つとして記念撮影が挙げられるが、初対面の相手に記念撮影を申し込むのは難しい。本システムでは、顔認識カメラを用いることで初対面の人間同士の記念撮影を行いコミュニケーション支援を行う。コミュニケーション支援として、本システムが人間に対して話題を投げかけることで記念撮影を行った人間同士での会話を誘発させるコミュニケーションの円滑化や、本システムが異なる人間との記念撮影を要求することで新たな初対面の人物とのコミュニケーションの誘発を行う。 | ||
| 金平 卓也 | 1551034: 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: Characterization of UGM using the Ontology based on the Location Information of POI
Abstract: Recently, some map services allow users to create their own maps for sharing spatial information visually and intuitively. These maps created by ordinary users are called a user generated map (UGM). UGM is one of the digital contents in which various POIs (Point of Interests) related to a certain topic are summarized in one map. These maps are useful information source when we visited on unvisited region. However, most of them are embedded into a home page, and it is hard to find a suitable map for a user's demand. Although Google started a gallery service of these maps from 2014, it is still difficult to find the desired map because each map does not include adequate and enough meta information. Since these are assumed to be embedded into a home page, meta information such as a topic of the map is usually described on the web (not on the map). In addition, names of POIs is often written simply without any category information. This paper aims to solve this problem of UGM by characterizing these maps with utilizing the ontology. The ontology is estimated based on the position information of POIs registered in the UGM. By characterization, it will be easy to find the UGM according to the user demand. In this presentation, we will describe the approach for characterizing UGM. Language of the presentation: Japanese 発表題目: POIの位置情報に基づくオントロジーを用いたUGMの特徴付け 発表概要: 近年、地図サービスでは、ユーザが視覚的かつ直感的に空間情報を共有するために独自のマップを作成することができる。これらの一般ユーザが作成したマップはUser Generated Map(UGM)と呼ばれている。UGMは特定のテーマに関連する様々なPOI(Point of Interest)が1つのマップにまとめられているデジタルコンテンツの1つである。このようなマップは、私たちが未訪問の地域を訪れた際の有用な情報源となる。しかし、ほとんどのUGMはホームページに埋め込まれており、ユーザの求める情報に適したマップを検索することが困難である。2014年にGoogleがUGMの検索することができるサービスを開始したが、各マップに適切かつ十分なメタ情報が含まれていないため、目的のマップを見つけることが困難な状況にある。UGMはホームページに埋め込まれることを想定されており、マップのテーマに関連するメタ情報は通常Webページ上に掲載されている。また、POIの名称の多くはカテゴリ情報がなくシンプルに書かれている。そこで、本研究ではオントロジーを用いてマップを特徴付けることにより、この問題を解決することを目的とする。オントロジーはUGMに登録されているPOIの位置情報に基づいて推定する。特徴付けを行うことでユーザの目的の情報が掲載されたUGMを発見することが容易となる。本発表では、UGMの特徴付けのアプローチについて述べる。 | ||
| 中田 彩 | 1551067: M, 1回目発表 | 環境知能学 |
|
title: Histograms of Oriented Gradients for Human Detection abstract: We study the question of feature sets for robust visual ob- ject recognition, adopting linear SVM based human detec- tion as a test case. After reviewing existing edge and gra- dient based descriptors, we show experimentally that grids of Histograms of Oriented Gradient (HOG) descriptors sig- nificantly outperform existing feature sets for human detec- tion. We study the influence of each stage of the computation on performance, concluding that fine-scale gradients, fine orientation binning, relatively coarse spatial binning, and high-quality local contrast normalization in overlapping de- scriptor blocks are all important for good results. The new approach gives near-perfect separation on the original MIT pedestrian database, so we introduce a more challenging dataset containing over 1800 annotated human images with a large range of pose variations and backgrounds. language of the presentation: Japanese | ||
| 西本 慎之介 | 1551074: M, 1回目発表 | 自然言語処理学 |
|
title: Detecting stance in tweets by using sentiment analysis
abstract: Sentiment Analysis is an important research topic in NLP. Sentiment Analysis is defined as determining whether the author’s attitude is positive or negative on a given text. However, it is more useful to find the opinion towards specific targets and positions than simply determining whether the author has a positive or negative attitude in general. For this reason, I deal with the detection of stances in tweets on my research. I define stance detection to mean automatically determining from text whether the author is in favor of the given target, against the given target, or whether neither inference is likely. Because of the fast changes in the political theme on the Internet, it often happens that the number of data is limited. Therefore, it is necessary to use external resources. In my presentation, the application of different methods of sentiment analysis and external resources will be shown. language of the presentation: Japanese | ||
| 前田 若菜 | 1551100: M, 1回目発表 | 知能コミュニケーション |
|
title: Privacy protection for a nonstructurized text data in social media using k-anonymization
abstract: It needs anonymization to open nonstructurized text data wihich includes personal information. Conventional methods make only a proper noun anonymous, therefore an anonymized phrase is able to estimated from un-anonymized characters. For solving it, this study suggests a method that a character N-gram phrase is anonymized based on k-anonymity. language of the presentation: Japanese | ||
| 須藤 広大 | 1551055: M, 1回目発表 | 自然言語処理学 |
|
Title:Relation Extraction with Distant Supervision Abstract:Recently,Relation Extracion is invaluable for arugment knowledge base that is useful for search or another application.Distant Supervision is useful approach for relation extraction.This method exceed previous method like Supervised Learning, Unsupervised Learning, Semisupervised Learning.So,I introduce Distant Supervision method for relation extraction.and,then propose extention of distant supervision and using Word Embedding to filtering relation extraction for get more acculate relation. Language of the presentation: Japanese | ||
| 澤田 一磨 | 1551050: M, 1回目発表 | インタラクティブメディア設計学 |
|
title: An Introduction of Robust Random Dot Markers Abstract: Different types of markers for Augmented Reality (AR) have been proposed in the last decades. Each type of marker comes with its advantages and disadvantages. Choosing a proper marker type is an essential decision to system’s success before developing an Augmented Reality application. Random Dot Markers (RDM), a keypoint based approach for detecting and tracking based on a dot pattern, is also one of those markers developed for AR applications. The biggest advantage of RDM is to be able to detect the marker even a part of it is occluded. However, overall robustness of tracking suffers when there is a large difference between the reference and detected markers. To overcome this limitation, Robust Random Dot Markers (RRDM) has been introduced by Yang et.al. RRDM provides more robust tracking than RDM since it employs a stricter matching process between reference and detected markers. In my presentation, I am going to introduce RRDM in detail. language of the presentation: Japanese 発表題目: An Introduction of Robust Random Dot Markers 発表概要: 拡張現実感を目的としてこれまで様々なマーカが提案されている．各種マーカはそれぞれに利点欠点を持ち合わせており，適切なマーカを選択することで設計したい拡張現実感アプリケーションの構築が容易になる．マーカの１つとして、ランダムに配置した点の集合をマーカとするRandom Dot Markers (RDM)が提案されている。RDMには、必ずしもマーカである点の集合をすべて映さなくても機能するという強みがある。しかし、カメラによって検出される点の集合と参照マーカの点の集合とのズレが大きい場合に、トラッキングのロバスト性に課題を残していた。これを克服した技術としてRobust Random Dot Markers (RRDM)が報告されている。RRDMはRDMに比べて、より厳密に参照マーカとの照合を行うことでロバスト性を向上させている。本発表では、RRDMについて説明する。 | ||