ゼミナール発表

日時: 9月28日(水)4限(15:10-16:40)


会場: L1

司会: 南 裕樹
新垣 杏里 1551004: M, 2回目発表 情報基盤システム学 藤川 和利,安本 慶一,笠原 正治,新井 イスマイル,猪俣 敦夫
title: *** replace this part with the title of your talk IN ENGLISH ***
abstract: *** replace this part with the abstract of your talk IN ENGLISH ***
language of the presentation: *** English or Japanese (choose one) ***
 
古川 智也 1551095: M, 2回目発表 情報基盤システム学 藤川 和利,伊藤 実,笠原 正治,新井 イスマイル,猪俣 敦夫
title: An overhead reduction method of key exchage protocol from the Ring-LWE problem with feature of communication data
abstract: Public key cryptosystems such as Diffie-Helman key exchange will become comprise by realization of a quantum computer. Hence key exchange protocol with Ring-LWE is one of post quantum cryptography is studied. But this protocolソ s communication data is twice than present key exchange protocol. Thereby the communication data increase explosively environment to use a large number of device such as IoT(Internet of Things). In my study, I proposed an overhead reduction method of key exchage protocol from the Ring-LWE problem with feature of communication data.
language of the presentation: Japanese
 
神内 良太 1551053: M, 2回目発表 情報基盤システム学 藤川 和利,杉本 謙二,伊藤 実,新井 イスマイル,猪俣 敦夫
title: A study of impacts and countermeasures in output signals from IoT devices under DoS attack
abstract: Recently, various devices begin to connect the Internet owing to the spread of IoT(Internet of Things). Therefore it has become an urgent matter to establish new security countermeasures. But IoT device is not able to deal with some new incidents because of difference in performance or operation between IoT and conventional ICT systems. In some cases, stability of IoT device is more important than using data. Therefore we should consider the necessary from new viewpoint of security. In my study, I show the result of load test and impact on output signals under DoS attack, and I discuss countermeasures and cause.
language of the presentation: Japanese
 

会場: L2

司会: 大和 勇太
萩原 康平 1551078: M, 2回目発表 環境知能学 萩田 紀博,小笠原 司,神原 誠之
title: Stress factor estimation of passenger based on the driving scene learning from a vehicle camera in autonomous driving
abstract: This presentation presents an approach for estimation of the passenger stress target when they ride autonomous vehicle. In recent years, autonomous car has become a hot topic beyond the car company. However, state of the art algorithms for autonomous vehicles mainly focus on collision free path planning and safe control for obstacle avoidance. So, our laboratory tried to make people comfortable while they riding the autonomous vheicle by our proposed method. But, these proposed approach can't adapt to dynamic environment and present information, change their behavior in real time recognition. In this paper, we propose a method for estimating the passenger stress target in a dynamic environment based on the driving scene learning from a image of the car-mounted camera, and develope a system that helps to improve the comfort of the passenger.
language of the presentation: Japanese
 
EL HAFI LOTFI 1461207: D, 中間発表 ロボティクス 小笠原 司,加藤 博一,高松 淳,丁 明
title: Gaze tracking using corneal images captured by a single wearable camera
abstract: The presentation will introduce a method to estimate the gaze direction using images of the eye captured by a single camera. The purpose is to develop wearable devices that enable intuitive eye-based interactions and applications. Indeed, camera-based solutions, as opposed to commercially available infrared-based ones, allow wearable devices to not only obtain natural user responses from eye movements, but also scene images reflected on the cornea, without the need for additional sensors. The proposed method relies on a model approach to evaluate the gaze direction and does not require a frontal camera to capture scene information, making it more socially acceptable if embedded in a glasses-shaped device. Moreover, recent development in high-sensitivity camera sensors allows to consider the proposed method even in low-light condition. Finally, experimental results using a prototype wearable device demonstrate the potential of the proposed method solely based on cornea images captured from a single camera.
language of the presentation: English
 
伊藤 淳 1551011: M, 2回目発表 ロボティクス 小笠原 司,加藤 博一,高松 淳,丁 明
title: Grasp motion analysis in daily life
In this study, we develop a wearable device for analyzing the grasp motion in daily life. For recording the grasp motion in daily life, we wear the motion sensors and the EMG sensors on the forearms, which can measure the joint angles and EMG signals simultaneously without restriction . In experiment, we confirmed that the developed device can measure the necessary data while grasping various objects.
language of the presentation: Japanese
発表題目:日常生活における把持動作の分析
発表概要: 本研究では,日常生活における人の手の動作を分析するために,ウェアラブルデバイスする.日常生活の把持を記録するため,拘束性の低いモーションセンサと筋電位センサを前腕に装着し,関節角度と筋電位データを同時に取得する.また,実験で複数の物体を把持しデバイスが分析の自動化に必要なデータを取得可能なことを確認した.
 
松浦 亮太 1551102: M, 2回目発表 ロボティクス 小笠原 司,加藤 博一,高松 淳,丁 明
title: Video summarization to support review of eating habits
abstract: Evaluation of eating habits is important to improve quality of life and health care. In this research, we propose a system to assist users in the review of their eating habits. From a video of a user having a meal, the proposed system recognizes the specific parts when the user is actually eating. To extract the eating segments from the video, we follow three different approaches: hand-head relation, clustering, and template matching. The training data is obtained through user interaction. Finally, we propose video summarization as the output of our system.
language of the presentation: Japanese
発表題目: 食生活の振り返り支援のための映像要約
発表概要: 健康管理や生活の質の向上のために,食生活を評価し,改善することは重要である.そこで本研究では,食事の様子を撮影し,ユーザが容易に食生活を振り返ることができるシステムの構築を目指す.本発表では,ユーザとのインタラクションを通じ,少量の学習データから学習された,撮影動画に特化した認識システムの可能性,およびそれを用いた映像要約について模索する.
 

会場: L3

司会: Juntao Gao
平部 裕子 1561017: D, 中間発表 ユビキタスコンピューティングシステム 安本 慶一,中村 哲,荒川 豊,諏訪 博彦,藤本 まなと
title: Recognizing user context through touch behavior analysis of a smartphone
abstract: There are many studies on user context recognition that utilize various sensors in a smartphone. However, most of them focus on a physical context such as behavior and posture. In order to realize more sophisticated context-aware services, we are trying to recognize more detailed human-related context such as profile information (e.g. dominant hand, operation skill, and age) and emotion. For this challenge, we focus on user's touch behavior while using a smartphone. In master thesis, we developed a system for obtaining all the touch behavior on any application in Android. With this system, in this study, we propose a method to estimate a user profile, especially operation form which means the combination of user hand and finger. We show that our proposed method achieves 96.46% accuracy for 4 operation forms, through 9 subjects' data.
language of the presentation: Japanese
 
柿木 研人 1551029: M, 2回目発表 ユビキタスコンピューティングシステム 安本 慶一,中村 哲,荒川 豊,諏訪 博彦,藤本 まなと

title: The relationship between Volatility Index and topic in stock forum

abstract: In finance, it is useful to predict various stock index. We focus on Volatility Index,that is called "Fear index". We hypothesize that Volatility Index and investor's information from social media have relationship.As an experiment, we analysed correlation and regression between classified topics from Yahoo! stock forum and Volatility Index.

As a result, we indicated the possibility of predicting Volatility Index .


インターネット株式掲示板における話題とVI指数の関係

ファイナンスにおいて,株式リターンや,ボラティリティだけでなく,様々な株式指標を予測することは有用である.我々は,恐怖指数と呼ばれ,投資家の不安を表す株式指標であるVI指数に着目し,ソーシャルメディアを分析することによって得られる投資家の情報と密接に関連のある指標であるという仮説を立てる.我々は,VI指数と話題の関係を明らかにする実験としてYahoo! JAPANの株式掲示板から取得したメッセージをトピック分類し,VI指数と相関分析,及び重回帰分析を行った.結果として,トピックの投稿率の変動によるVI指数の予測可能性を示した.

 
金平 卓也 1551034: M, 2回目発表 ユビキタスコンピューティングシステム 安本 慶一,中村 哲,荒川 豊,諏訪 博彦,藤本 まなと
Title: Similarity Search System for My Map in Unvisited Region
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 My Map. My Map 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 My Map by characterizing these maps with utilizing the ontology. The ontology is estimated based on the position information of POIs registered in the My Map. By characterization, it will be easy to find the My Map according to the user demand. In this presentation, we will describe the approach for characterizing My Map.
Language of the presentation: Japanese
発表題目: 未訪問地域の類似マイマップ検索システムの提案と実装
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 マイマップと呼ばれている。マイマップ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、ほとんどのマイマップ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がマイマップの検索を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。マイマップ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、本研究ではカテゴリベクトル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。カテゴリベクトルはマイマップに登録されている各POI名称および位置情報に基づいて生成する。カテゴリベクトルによ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マイマップを発見することが容易となる。本発表では、提案する類似マイマップ検索システムおよびカテゴリベクトル生成方法について述べる。
 
中川 愛梨 1551066: M, 2回目発表 ユビキタスコンピューティングシステム 安本 慶一,中村 哲,荒川 豊,諏訪 博彦,藤本 まなと

title: Real-time activity recognition system in a smart home

abstract: Recently, there are many studies on automatic recognition of activities of daily living (ADL) to provide various services such as elderly monitoring, intelligent concierge, and health support. In particular, real-time ADL recognition is essential to realize an intelligent concierge service since the service needs to know user's next activity for its support. Our previous work already achieved in-home living activity recognition with 91% of accuracy. However, it takes 5 minutes to recognize each activity because of coarse granularity (1/30 Hz) of collected sensor data. In this study, to achieve real-time living activity recognition, we propose a sensing system that can collect finer granularity (1 Hz) of the sensor data, and a machine learning-based recognition method that use new features.

language of the presentation: Japanese