| 若元 亮樹 | 1451120: M, 1回目発表 | ソフトウェア工学 |
| title: Toward Identifying Future Committers based on Developers' Communications abstract: Open Source Software committer is a role of developer who has permission to commit edited source code (e.g., patches) to version control system of his or her projects. An OSS project gives such permission to a very limited number of capable developers. However, though most of OSS projects do not have guideline to apprise developers, current committers would select based on their different criteria.Previous studies investigated metrics (e.g., patch submission, comment submission) to make a guideline to identify a capable developer. We consider that current committers are not only their frequency but they focus on what developers contributed to their projects. In this study, we investigate how developers’ contributions would evolve before becoming a committer, and we would like to make a new guideline to identify a future committer. language of the presentation: Japanese | ||
| 森 仁志 | 1451111: M, 1回目発表 | ソフトウェア設計学 |
|
title: Paper introduction of ”Evaluating Source Code Summarization Techniques:Replication and Expansion” abstract:During software evolution a developer must investigate source code to locate then understand the entities that must be modified to complete a change task. To help developers in this task, Haiduc et al. proposed text summarization based approaches to the automatic generation of class and method summaries, and via a study of four developers, they evaluated source code summaries generated using their techniques. Authors proposed new topic modeling based approach to source code summarization via a study of 14 developers, and evaluated source code summaries generated using the proposed technique. language of the presentation: Japanese | ||
| 川崎 真司 | 1451037: M, 1回目発表 | ディペンダブルシステム学 |
|
title:Soft Error Sensitivity Evaluation of Microprocessors by Multilevel Emulation-Based Fault Injection
In order to satisfy reliability requirements, understanding the behaviour of a circuit with failure is important. Fault injection is widely used for such a purpose. Since software approaches are quite slow, emulation-based approaches have been accepted to accelerate the fault injection process. These methods use FPGAs to prototype the Circuit Under Test (CUT) and support the fault injection mechanisms. In my presentation, I will introduce an emulation-based approach that can efficiently emulate the real delays of CUT produced for the target ASIC technology on FPGA. I will also present my research objective and plan. language of the presentation: Japanese | ||
| 里中 沙矢香 | 1451053: M, 1回目発表 | ディペンダブルシステム学 |
|
title:Isometric Test Compression with Low Toggling Activity
abstract:Modern complex designs demand more reliable and effective testing with as much reductions in storage requirements as possible. Many diverse test data compression techniques are widely recognized as essential to reduce the overall test cost. In this presentation, I introduce a novel test data compression scheme that synergistically elevates compression ratios to values typically unachievable through conventional reseeding-based solutions. My research objective and plan are also presented. language of the presentation: Japanese | ||
| AKPA AKPRO ELDER HIPPOCRATE | 1451122: 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: Smartphone-based Calorie Estimation System
abstract: People across the world are becoming more and more interested in watching their weight, eating healthy and avoiding obesity. As overweight and obesity are major risk to human health, a system that can measure calories and nutrition in every day meal can be very useful. Recently, calorie and nutrition intake measurement systems for overweight and obesity treatment have been the focus of many studies. In this presentation, we propose a food calorie and nutrition estimation system (ICE CREAM) that can help obese patients, dietitians and ordinary people to easily measure and manage daily food intake. Our proposed system is mainly built on smartphone's sensors, which is practically carried by users all the time. By using a food picture and user gestures on the food, our system deals with the challenging volume measurement step of existing calorie estimation system. The experimental results of our classification methods demonstrate our results with respect to accuracy and reliability. language of the presentation: English | ||
| CHAN KAR LONG | 1351203: M, 2回目発表 | 飯田 元，加藤 博一，市川 昊平 |
|
title: A Hybrid game contents streaming method to improve graphic quality delivered by cloud gaming.
abstract: In gaming industry, cloud gaming is a new form of gaming service on trend trying to serve millions of players around the world with novel gaming experience. In this research, the proposed hybrid game contents streaming method aims at improving graphic quality delivered from cloud gaming. Game contents streamed from cloud server are splitted into two parts, as one part is streamed as video sequence. Another part, which is streamed as graphic commands, is for rendering job to be performed at the client side. The locally rendered game contents are then displayed accordingly together with the contents streamed as video sequence. By taking the advantage of distributing rendering operation, the combined computing power is expected to produce better graphic quality. Graphic quality delivered from this hybrid model will be compared with traditional video streaming method used by other existing cloud gaming platforms. language of the presentation:English | ||
| 瀬戸 将志 | 1451061: M, 1回目発表 | インタラクティブメディア設計学 |
|
title: The research of the estimation of camera localization and the correction of image distortion for AR
abstract:Augmented Reality (AR), requires the simultaneous correction of rolling shutter image distortion caused by camera or object motion and estimation of camera localization. But, there has not been any research about these two problems together, only about the correction of image distortion. Therefore, this study analyzes the conventional studies and proposes an approach for AR language of presentation: Japanese 発表題目: ARのための高精度な自己位置推定と画像の歪み補正に関する研究 発表概要: AR、拡張現実感では観測した映像に情報を重ね合わせるため、カメラ・被写体運動によって生じる画像のローリングシャッター歪み補正とカメラの自己位置推定を同時に行う必要がある。しかし、従来研究では歪み補正のみを目的としており、これらの相関関係に着目した研究はなかった。そこで本研究では従来研究の手法を分析するとともに、ARへの応用のためのアプローチ方法を提案する。 | ||
| ROMPAPAS DAMIEN CONSTANTINE | 1451127: M, 1回目発表 | インタラクティブメディア設計学 |
|
Title: Dynamic Augmented Reality X-Ray on Google Glass Handheld devices and smart glasses are well equipped to build a knowledge base of useful spatial information. Presenting this information in a way that can help people achieve tasks such as navigating in unknown indoor or urban environments is desirable. For this purpose, Augmented Reality X-Ray is an intuitive visualization technique. Previously, AR X-Ray systems either utilize only static images as source data, or a stationary camera. We present a system that grants its users rotational freedom, processing live video data. The system uses sensory capabilities of two Google Glasses to provide one user (watchman) with an AR X-Ray of the second user’s (scout) observed scene. Language of Presentation: English | ||
| 太田 翔平 | 1451024: M, 1回目発表 | 環境知能学 |
|
title: Stress measurement of Autonomouse Vehicles passenger by physiological indices
abstract:This paper investigates stressors of driver of an autonomous vehicle by measuring physiological indices. Technologies of autonomous vehicles have been researched in order to reduce a traffic accident and stresses by traffic jam. Government aims to realize Semi-autonomous vehicle which drives by a human only in case of emergency situation by early 2020’s. In the semi-autonomous driving, since a driver should watches out an environment around car to avoid accidents, the driver feels stresses. This paper investigates stressors which a driver feels under semi-autonomous driving by measuring physiological indexes. language of the presentation: Japanese 発表題目:生理指標を用いた自動走行車両の搭乗者のストレスに関する研究 発表概要: 近年，交通事故の低減，渋滞緩和を目的として，自動車の自動走行化に向けての研究が盛んに行われており，2020年代前半に緊急時のみ搭乗者が対応する準自動走行の実現が目標として掲げられている。 自動走行の普及には，搭乗者の快適性への配慮も重要な課題であるが，準自動走行では，搭乗者は常に運転状況を把握し，緊急回避行動に備える必要があるため，その心構えの継続からストレスを感じ，快適性が低下することが容易に想像される。 本研究ではこの問題を解決するため，走行状況と生理情報の観測に基づいた心理分析により，準自動走行状態において搭乗者が感じるストレスの要因の解明を目指す。 そのために，準自動走行車両を想定し，助手席搭乗者のストレスを生理指標を用いて測定することで，どのようなストレスを受けるかを調査した． | ||
| �矍澄々�太郎 | 1451063: M, 1回目発表 | 環境知能学 |
|
title: Method of acquisition and usage of user information using interactive sensing robot
A decrease in utterance opportunities of elderly people living alone is a social problem. The problem increases the number of people who has depression and dementia. Although there are people who helps them by listening and counseling, a shortage of volunteer workers is a problem. To solve the problem, there are many researches of conversation robots. In this research, we aim to develop a utterance system using social comments and a interactive sensing which collects user's information by talking with robots. Japanese | ||
| 大西 真輝 | 1451026: M, 1回目発表 | 自然言語処理学 |
|
title: Complex Coreference Resolution using Big Dependency Knowledge
abstract: Coreference resolution is the task of finding all expressions that refer to the same entity in a text. In this study, we examine complex case of them which traditional linguistic constraints on coreference as well as commonly-used resolution heuristics are not useful. We use a knowledge-rich approach to this task by a lot of dependency knowledge from annotated Gigaword. language of the presentation: Japanese 発表題目: 依存構造知識を用いた複雑な共参照問題の解析 発表概要: 共参照解析とは，文書中の表現を現実世界の同一な実体ごとに分類するタスクである．本研究では特に，自然言語処理で使用される伝統的な技法では解くことが難しいとされている，複雑な共参照問題について取り扱う．我々は，メタデータが付与された大規模なGigawordコーパスを，依存構造知識として利用することでこのタスクに取りんだ． | ||
| 前川 拓也 | 1451094: M, 1回目発表 | ネットワークシステム学 |
|
title:Efficiency Regularization of Parallel Transmission Line-Based Wireless Power Transmission Using Duplicated Lines
abstract:This paper proposes a wireless power transmission system using duplicated feeder lines. In wireless power transmission using parallel feeder line, the power transmission efficiency depends on the location of the secondary circuit due to the constant wave on the feeder line. The conventional method using reactance loading can regular constant wave, but is affected by secondary loading. We propose the method for efficiency regulation providing another feeder line that shifted 45 phase. In this method, power receiving side is equipped with two coils, and it is arranged to be powered from the line respectively. language of the presentation: Japanese | ||
| 西�� 優弥 | 1451080: M, 1回目発表 | 視覚情報メディア |
|
title: Indirect Augmented Reality with variable viewpoints
abstract; In mobile Augmented Reality (AR) applications using a smartphone, the geometric registration to accurately place virtual objects is one of the essential requirements. In particular, the jitter of virtual objects between frames causes the reduction of the quality of AR images. Conventionally, to solve this problem, a Indirect AR approach has been proposed. This method can give experiments of AR to users without the jitter because this method uses pre-captured omnidirectional images on which virtual objects are superimposed in advance. However, one of the problems for this method is that it is impossible to see the scene with variable viewpoints. Then, this study proposes a Indirect AR method with variable viewpoints. In this presentation, I introduce a prototype system based on a conventional method and present experiments using the prototype system and future prospects. language of the presentation: Japanese 発表題目: 視点移動可能な事前生成型拡張現実感システム 発表概要: スマートフォン等を用いたモバイル拡張現実感(Augmented Reality:AR)アプリケーションにおいて，仮想物体を意図通りの場所に正確に配置するための幾何学的整合性は満たすべき条件の一つである．特に，フレーム間で仮想物体が細かく揺れ動くジッタはAR映像の品質を低下させる大きな要因となっている．従来，この問題を解決する手法に事前生成型拡張現実感がある．この手法は，実環境を事前に撮影した全方位画像に予め高品位に仮想物体を重畳・合成し，利用時にその全方位画像からユーザの向きに応じた方向を切り出して提示するものであり，ユーザはジッタのないAR映像の体験が可能となる．しかし，この手法の問題点の一つに，事前に撮影された地点での景観しか提示できないことが挙げられる．そこで，本研究では，ユーザが任意の位置に移動してもその地点での高品位なAR映像を提示することができる視点移動可能な事前生成型拡張現実感システムを提案する．本発表では，予備実験として従来手法に基づいて試作したシステムの紹介を行い，その考察と今後の展望について述べる． | ||
| 宮本 拓弥 | 1451105: M, 1回目発表 | 視覚情報メディア |
|
title: Paper introduction of ”LSD SLAM : Large Scale Direct Monocular SLAM”
abstract: These days, there have become increasing research topics because of unmanned aerial vehicles(UAVs) movement in the place which people can't enter in. When MAV moves from start point to goal point, it is necessary to get position information. To get the position information, SLAM(Simultaneously Localization and Mapping) has been proposed. This method simultaneously localizes its own position and maps its environment. However, SLAM has a problem about accumulated error. In my presentation, I will introduce LSD SLAM(Large Scale Direct Monocular SLAM). And then, I will analyze the execution result of SLAM which is open in GitHub. Finally, I will state about the problem and future plan. language of the presentation: Japanese 発表題目:論文紹介"LSD SLAM : Large Scale Direct Monocular SLAM" 発表概要: 近年, 火災や自然災害など, 人の進入が困難な場所で無人ロボットや無人飛行機を自律的に移動させる研究が盛んに行われている. 自律的にロボットを動作させるためには, 現在地から目的地へ移動する過程で位置情報が必要になる. 位置情報を取得するために, 動画から周辺環境の三次元形状復元とカメラ位置姿勢を推定するトラッキングを組み合わせたSLAM(Simultaneously Localization and Mapping)があるが, 長時間動作させると誤差が蓄積される問題がある. 本発表では, 蓄積誤差を減らすために, 平行移動, 回転移動, スケール変化を考慮したトラッキングとノイズを含むデプスの影響を確率的に排除する手法を用いたLSD SLAMについて紹介する. また, GitHubで公開されているLSD SLAMのプログラムの実行結果を分析し, 問題点と今後の展望について述べる. | ||
| 杉山 享志朗 | 1451060: M, 1回目発表 | 知能コミュニケーション |
|
title: Cross-Language Adaptation for Question Answering System
abstract: Constructing question answering system (QA system) requires many linguistic resources. Therefore, it is difficult to create systems for languages with few resources. In this study, I propose cross-language adaptation of an English QA system to construct systems in such languages. This proposed method may make it possible to build a system with less cost than make new resources. language of the presentation: Japanese 発表題目: 質問応答システムの言語適応 発表概要: 質問応答システムの構築には大量の言語資源が必要とされ、その資源の作成に大きなコストがかかるため、様々な言語での質問応答システムの構築を困難にしている。本研究では、資源が乏しい言語での質問応答システム構築のため、資源が豊富な言語である英語で作成された質問応答システムを他言語へ適応する手法を提案する。これにより、新たな資源を作成するよりも少ないコストでのシステム構築を可能とすることを目標とする。 | ||