| 桶田 真吾 | M, 2回目発表 | サイバネティクス・リアリティ工学 | 清川 清, 加藤 博一, 佐藤 智和（客員）, 酒田 信親 |
|
title: Toward More Believable VR by Smooth Transition between Real and Virtual Environments via Omnidirectional Video
abstract: In conventional virtual reality systems, users usually do not perceive and recognize the experience as reality. For example, users of a virtual disaster simulator know, consciously or unconsciously, that the presented disaster is not real, which inherently limits the training effect. To make the virtual experience more believable, we propose a novel real-virtual transiton technique that preserves the sense of “conviction about reality” in a virtual environment. This is realized by spatio-temporal smooth transition from the real environment to the virtual environment with omnidirectional video captured in advance at the user’s position. Our technique requires less preparation cost and presents a more believable experience compared to existing transition techniques using a handmade 3D replica of the real environment. In this presentation, reported are the concept of our technique, a prototype system and a preliminary user study that has shown the effectiveness of the technique. language of the presentation: Japanese | |||
| 常木 健介 | M, 2回目発表 | ソフトウェア設計学 | 飯田 元, 松本 健一, 市川 昊平, 崔 恩瀞 |
|
title: Investigation of the nature of code clones aimed at analysis support during Code Clone maintenance work
abstract: A code clone refers to a code fragment having the same or similar portion present in the source code. In general, code clones can be a factor in lowering maintainability of software, so developers need to perform maintenance work (eg, simultaneous modification, aggregation) on detected code clones. However, when many code clones are detected from large-scale software, it is difficult for the developer to find the code clone to be subjected to maintenance work from among them. Therefore, it is necessary to support the analysis work after detecting the code clone, and support the maintenance work of the developer. For example, by sequencing maintenance work on many detected code clones, the developer can focus on high priority clones. In this research, in order to support analysis during maintenance work, focusing on the positional relationship of code clone pairs, investigating the nature of code clones that should be handled preferentially as maintenance work from maintenance contents and maintenance frequency, To clarify. language of the presentation:Japanese | |||
| 中西 駿太 | M, 2回目発表 | ソフトウェア設計学 | 飯田 元, 松本 健一, 市川 昊平, 崔 恩瀞 |
|
title: Towards Understanding of OSS Development Activities by Topic Extraction
abstract: To reduce software development cost, Open Source software (OSS) is frequently used in software development companies. They want to know about target OSS development activity, but it is difficult. It is cited as a disadvantage of OSS. therefore in this research, I focused on discussion in OSS development. I aim to present the characteristics of development activities by extracting topics of discussion among OSS developers. In this presentation, I describe the results of examining what kind of topics are discussed in the issue management system. language of the presentation: Japanese | |||
| 湯口 彰重 | D, 中間発表 | ロボティクス | 小笠原 司, 中村 哲, 高松 淳, 丁 明, Gustavo Garcia |
|
title: Human-like gaze behaviors on a humanoid communication robot using its eyes
abstract: Android robots are used as communication robots due to their human-like appearance and their ability to generate gaze behaviors using their eyes. To emulate human communication, it is important for android robots to have human-like gaze behaviors. In this work, we present the strategy to achieve human-like gaze behaviors on a humanoid communication robot using its eyes. First, we propose a gaze calibration method for an android robot as a communication robot using a single camera as the gaze direction target in order to accurately adjust the android robot's gaze. Second, we present the measurement of a humans' eye movement behaviors such as eye contact or looking away and blinking to emulate the behaviors on an android robot using its eyeballs. Third, we present a method for estimating an interaction partner's gazed object from a service robot view using a relationship between a human's face orientation and surrounding objects' position to understand what an interaction partner look at from the robot's view. Finally, we merge each step to generate human-like gaze behaviors on a humanoid communication robot using its eyes. language of the presentation: Japanese | |||
| 中村 匠 | M, 2回目発表 | ロボティクス | 小笠原 司, 中村 哲, 高松 淳, 丁 明, Gustavo Garcia |
|
title: Robot operation of home appliances using a knowledge database
abstract: In a home environment, operating home appliances is very common but it is a difficult task for robots because of the various functions that home appliances have. In this research, we propose a robot system that uses a knowledge database and a GUI. In the GUI, humans support the robots' learning of the operating methods of home appliances and store the flow of these operations into a knowledge database. The robot uses RGB-D sensors to recognize home appliances and their operation, stores this information on the knowledge database, and operates the home appliances. In this presentation, we show the current progress of the system implementation and discuss the feasibility of our approach to facilitate the operation of home appliances using constantly learning robot systems. language of the presentation: Japanese | |||
| 秋山 解 | D, 中間発表 | 光メディアインタフェース | 向川 康博☆, 中村 哲, 金出 武雄, 舩冨 卓哉, 伍 洋 |
|
title: Hand Tracking with Wearable Depth Cameras and Convolutional Neural Networks
abstract: In this research, hand tracking from depth images is studied from the aspects of camera layout and algorithm. Depth images for hand tracking are captured typically from cameras fixed in the environment or mounted on the user's head. In these cases, it is highly possible that the hand is occluded by the body or deviated from the camera's view depending on the user's pose, which is fatal for hand tracking. A wearable camera mounted on the wrist, which is the closest part to hand, would improve view and become a flexible option for hand tracking regardless of pose. Hand tracking algorithm was developed under the public challenge. The provided dataset is not recorded from wrist, but simply transferring the CNN model trained for the challenge to wrist-mounted cameras takes effect to specific extent. language of the presentation: Japanese | |||