| 大石 陽波 | 1551020: M, 1回目発表 | 環境知能学 |
|
title: [Paper Introduction] Mental Workload and Task Performance in Peer-Based Human-Robot Teams
abstract: Successful human-robot (H-R) teams working with direct interaction and close coupling will have relationships that vary, depending on whether or not collaboration is prioritized and how collaboration changes the human mental workload and the team’s task performance. Modeling of representative functions can provide predictions of changes in mental workload and the potential impact on team performance. The presented research focuses on modeling and quantifying mental workload for H-R teams in which team members have some individually assigned responsibilities but must also make joint decisions with a teammate. IMPRINT Pro, a discrete event simulation modeling tool created by the U.S. Army Research Laboratory, was used to model human-human (H-H) and H-R teams completing a reconnaissance task in a building. This research evaluated H-H and H-R teams completing the same reconnaissance tasks. language of the presentation: Japanese | ||
| 櫛田 貴弘 | 1551039: M, 1回目発表 | 光メディアインタフェース |
|
title: Measurement of reflection properties by using scattered light
abstract: Measuring reflection properties from real objects is necessary for creating realistic CG. But it takes huge amount of time becaue it is required to measure from every incident and reflected angle. To overcome this problem, we measure the reflection property by measuring the scattered light of the reflection and reconstructing the distribution of light. language of the presentation: Japanese 発表題目: 散乱を利用した物体の反射特性の計測 発表概要:写実的なCGを生成するために，実物体の表面反射特性を計測し，それを利用する方法がある．そのためには，物体の表面反射特性の計測が必要であり，実物体からこれを計測することが求められている．反射特性の計測は通常，光の入射角と反射角の膨大な組み合わせについて計測する必要があり，容易に計測することはできない．本研究では，光の散乱を利用して反射光を観測し，空間中の反射光の分布を求めることにより，物体の反射特性を計測することを目指す． | ||
| 池田 大志 | 1551006: M, 1回目発表 | 自然言語処理学 |
|
title: An Encoder-decoder Model for Japanese Text Normalization
abstract: Microblog such as Twitter has become an important research topic in NLP. However, mircoblog texts are often written in an informal style and include many lexical variants. Current tools for Japanese word segmentation and POS tagging still perform poorly on microblog texts. A normalization task is very difficult to conduct in Japanese because there are no explicit boundaries between words and Japanese has three types of characters: Kanji, Hiragana and Katakana. To address this issue, we propose the method of cleansing noisy Japanese text using an encoder-decoder translation model. language of the presentation: Japanese | ||
| 寺崎 希 | 1551064: M, 1回目発表 | 視覚情報メディア |
|
title: Image-Based Rendering in the Gradient Domain
abstract: I will introduce a research paper titled “Image-Based Rendering in the Gradient Domain”. It was published by J.Kopf et al. in 2013. In this paper, they propose a image-based rendering algorithm for handling complex scenes that may include reflective surfaces. Their key contribution lies in treating the problem in the gradient domain. They use a standard technique to estimate scene depth, but assign depths to image gradients rather than pixels. A novel view is obtained by rendering the horizontal and vertical gradients, from which the final result is reconstructed through Poisson integration using an approximate solution as a data term. language of the presentation: Japanese | ||
| 秋山 諒 | 1551002: M, 1回目発表 | インタラクティブメディア設計学 |
|
title: Appearance Control Considering Reflectance and Environmental Light for Projector - Camera System
abstract: A lot of technologies that control appearance of physical surface by using a projector and a camera are implemented. One of them is applying a feedback system for calculating proper projection images. However, the method requires that an environmental light is constant. Therefore, we cannot apply it to most of real life such situation where the environmental light is changing inconstantly. In this study, we estimate reflectance and variation of environmental light simultaneously and control appearance under dynamical environment light. In this presentation, I will explain feedback system with projector camera that was used in conventional method, and show our estimation method. Additionally, I will report about current progress. language of the presentation: English | ||
| 今西 雅美 | 1551014: M, 1回目発表 | 生体医用画像 |
|
title:
Automated muscle Segmentation from 3D CT data of lower extremities with varing joint postures using hierarchical multi-atlas method abstract:Automated muscle Segmentation from 3D CT data of lower extremities with varing joint postures using hierarchical multi-atlas method has been proposed in my laboratory. When the patients lies on thier backs, their postures are difference because their bones is deformated or they have pain.But privious method hasn't considered this. So, I'll study the stable method to operate if their postures are diffrence. As preparing experiment, segmentation of only one side will be done.I'll divide a domein every bone. language of the presentation: Japanese | ||
| 諏訪部 開 | 1551056: M, 1回目発表 | 数理情報学 |
|
title: Collecting thought reports during mind wandering: toward neural decoding of subjective mental time
abstract: We spend half of our waking time with thoughts independent of what is going on around us. While such task-unrelated thoughts, mind wandering, can be a cause of error in external tasks, studies revealed functional effects of mental time during mind wandering on the adjustment of mental mood. Here, to reveal the neural substrates of mental time in task-unrelated thought, we aims at performing brain decoding of mental time from the brain activity during mind wandering. Before launching brain measurement experiments, we first conducted behavioral experiments to confirm the efficiency of collecting thought reports during mind wandering. By asking subjects to be engaged in a simple choice reaction time task and to report the contents of their mind wandering, we were able to collect around 10 reports of task-unrelated thought in a 40-minutes experiment. language of the presentation: Japanese | ||
| 能登 健太朗 | 1551075: M, 1回目発表 | 知能システム制御 |
|
title: Generalization of Jitter Compensation by Gain Switching Observer abstract: In networked control systems, due to its congestion, the length of transmission delay varies irregularly, which is called ”jitter.” In such systems, a gain switching observer via time-stamp information was proposed for state estimation with jitter compensation. This study aims to remedy a drawback of the conventional observer when multiple signals arrive at the same time. language of the presentation: Japanese | ||
| 本田 新 | 1551098: M, 1回目発表 | 数理情報学 |
|
title: Derivation of the learning curve of perceptron using statistical mechanics
abstract:Statistical mechanics method can apply to information problem such as the image restoration and neural network, and that the framework is called information statistical mechanics. Information statistical mechanics can analyze the perceptron learning algorithm. The perceptron learning algorithm is an on-line learning algorithm that the perceptron updates its connection weight to approach hopeful output (teacher's signal). Miyoshi extended the analysis to the ensemble learning and/or noisy cases. Among them, the learning curve(the learning process) has specific behavior when the teacher's signal has observation noise. In this study we derive the learning curve of perceptron with noisy teacher's signal using statistical mechanics, and analyze the behavior around the convergence point to explain why and how occurs it. language of the presentation: Japanese | ||
| 伊藤 淳 | 1551011: M, 1回目発表 | ロボティクス |
|
title: Analysis of the grasp using the multiple sensors in daily life
abstract: Understanding how humans grasp objects is important when performing grasps with a robot hand that performs like a human hand. Previous research utilizes a camera for grasp analysis. But the use of cameras poses privacy issues, and cameras cannot obtain certain information. In this study, we analyze the grasp state automatically using an existing taxonomy based on composite sensors values rather than camera data, namely inertia, myoelectric and contact sensors. language of the presentation: Japanese 発表題目:複合センサを用いた日常生活における把持動作の分析 発表概要:人と同等の把持動作が行えるロボットハンドを開発する上で，日常生活で人間がどのように把持動作を行っているかを分析することは重要である． これまでカメラを利用した把持動作の分析は行われているが，カメラで取得できない情報や，プライバシ保護に課題がある． 本研究では，慣性，筋電位，接触などカメラ以外の複数のセンサから取得した情報を利用することによって，把持動作を高精度に認識する手法を開発し，日常生活における把持動作を分析する． | ||
| 増田 健一 | 1551101: M, 1回目発表 | モバイルコンピューティング |
|
title: Study of Reservation-Based Road Pricing System to Reduce Traffic Congestion
abstract: Traffic congestion causes many problems, for example, an economic loss and environmental pollution. To reduce an influence of the traffic congestion, road pricing schemes have been proposed and implemented. Variable tolls based on traffic conditions are effective for high efficiency of road utilization, however, the users in the existing road pricing methods cannot know the prices to pay actually because of the difficulty of predicting traffic conditions. This research aims to provide comfortable traffic environments without traffic congestion by selling tickets to users in advance of using roads. In this presentation, I introduce the existing road pricing methods and those problems. After that, I explain my researching method for a reservation-based road pricing scheme. language of the presentation: Japanese | ||
| 新垣 杏里 | 1551004: M, 1回目発表 | 情報基盤システム学 |
|
title: Malware Detection
abstract: In recent years, with the rapid development of the Internet, frequency of occurrence and the propagation speed of malware has become a serious threat. It is difficult to deal wth unknown malware and 0-day attacks by conventional pattern matching method. Therefore, studies to detect using binary data or behaivor of malware is activily. So, I will introduce research on malware detection. language of the presentation:Japanese | ||
| 田村 洋輔 | 1551061: 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: Meal assist system by IoT device abstract: Suggestion of meal assist system by IoT device. It mainly deals with sensing data of human activity in kitchen and recognize it. language of the presentation: Japanese | ||
| 佐藤 敬済 | 1551048: M, 1回目発表 | ディペンダブルシステム学 |
|
title: (Paper Introduction) Burn-In time reduction based on parametric test analysis
abstract: Burn-In test is an effective test process to screen out early-life failures of VLSI circuits. However, burn-In test requires the expensive equipment and long test application time. Test data analysis provides an opportunity to reduce these costs. Burn-In test result can be predicted by machine learning using existing test result. In my presentation, I'm going to introduce an overview of related papers on test data analysis. language of the presentation: Japanese | ||
| 中川 愛梨 | 1551066: 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: Real-time activity recognition in a smart home
abstract: Recently, there are many studies on in-home living activities recognition which are expected to be applied to life support services, such as home appliance control for energy-saving and elderly monitoring. To realize these services, real-time activity recognition is essential. Our previous work already achieved in-home living activity recognition with 97% 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 発表題目: スマートホームにおけるリアルタイム行動認識 発表概要: 近年，屋内での生活行動を自動認識する研究が盛んに行われており，省エネ家電制御や高齢者見守りシステムなどの生活支援サービスへの応用が期待されている．それらのサービスの実現には，リアルタイムでの行動認識が必要となる．先行研究において，97%の精度で行動認識しているものの，認識まで５分のデータを必要としており，リアルタイムに対応できていない．本研究では，リアルタイム行動認識のために，１秒ごとにデータを収集するシステムと，新たな特徴量を使用する手法について提案する． | ||