The 9th COE Postdoctoral and Doctoral Researchers
Technical Presentation

Date: Wednesday, Dec. 22, 2004
Time: 13:30 - 14:30
Place: L1 Lecture Room
Language: English (Oral Presentation), English/Japanese (Question)
Chairperson: Osamu Tobe (Internet Architecture and Systems Lab. : D2),
Masanao Koeda (Robotics Lab. : D3)

Program (20 mins each: 15 mins presentation and 5 mins discussion)

  1. "Progress in audible (nornal) and inaudible speech recognition based on NAM microphones"
    Panikos Heracleous (Speech and Acoustics Laboratory : PD)
    パニコス イラクレウス (音情報処理学講座 : PD)

    [Abstract]
    Previously, we presented speaker-dependent experimental results of inaudible murmur (NAM) speech recognition with 93.5% word accuracy, which is a very promising result. In this work, we will introduce preliminary speaker-independent experimental results using training data of several speakers. The results are resonable and show the effectiveness of our approaches in speaker-independent case, too. The latest versions of NAM microphone willl be also introduced and some speaker-dependent experimental results under various environments will be given. Apart from speech recognition applications, NAM microphones can be also used in human-human communication for noise robustness puroposes. To evaluate the perceived speech, we investigate the use of objective measures for speech intelligibility and speech quality.
    Top
  2. "Fast-Convergence Blind Source Separation Combining ICA and Beamforming"
    Tsuyoki Nishikawa (Speech and Acoustics Laboratory : D3)
    西川 剛樹 (音情報処理学講座 : D3)

    [Abstract]
    We propose a new blind source separation (BSS) algorithm for multiple source signals. In the proposed algorithm, independent component analysis (ICA) and beamforming are combined to resolve the slow-convergence problem through optimization in ICA. The proposed method consists of the following three parts: (a) frequency-domain ICA with direction-of-arrival (DOA) estimation, (b) null beamforming based on the estimated DOA, and (c) integration of (a) and (b) based on the algorithm diversity in both iteration and frequency domain. The separation matrix obtained by ICA is temporally substituted by the matrix based on null beamforming through iterative optimization, and the temporal alternation between ICA and beamforming can realize fast- and high-convergence optimization.
    Top
  3. "Estimation of Focus of Attention of Multiple People for Video Conferencing"
    Kentaro Takemura (Robotics Laboratory : D2)
    竹村 憲太郎 (ロボティクス講座 : D2)

    [Abstract]
    In this presentation, we propose a method to estimate the focus of attention of multiple people. It utilizes the gaze information as 3D vectors and estimates the position of focus of attention defined as the intersection of the multiple vectors. As the results, the position of the focus of attention can be represented as a continuous 3D position unlike other researches where only discrete objects can be the focus of attention. As experiments, the focus of attention at word-chain game is estimated and the feasibility of the method is confirmed. We applied the proposed method to video conferencing, and confirmed the feasibility through experiments.
    Top
  4. "Evaluating the Risk of Information Leakage in Security-Sensitive Software Process"
    Yuichiro Kanzaki (Software Engineering Laboratory : D2)
    神崎 雄一郎 (ソフトウェア工学講座 : D2)

    [Abstract]
    I will be talking about a method to evaluate the risk of information leakage in a software process for security-sensitive applications. A software process is modeled as a series of sub-processes, each of which produces new work products from input products. Since a process is conducted usually by multiple developers, knowledge of work products is shared among the developers. Through the collaboration, a developer may tell others the knowledge of products that are not related to the process. We capture the transfer of such irrelevant product knowledge as the information leakage in a software process. In this presentation, I first formulate the problem of information leakage by introducing a formal software process model. Then, I propose a method to derive the probability that each developer d knows each work product p at a given process of software development.
    Top
  5. "ALTAIR: Automatic user location system using Active IR-tag"
    Muneyuki Sakata (Image Processing Laboratory : D2)
    坂田 宗之 (像情報処理学 : D2)

    [Abstract]
    In ubiquitous computing environment, user location is one of the most important information. Global Positioning System (GPS) is most powerful tool in outdoor use, but it can not be used in indoors. In this presentation, I describe our user location system for indoors named ALTAIR. ALTAIR is the system that detect users' position and identify users automatically. The system uses wireless network and combination of IR(Infra-Red)-tags and wide angle cameras.
    Top

21st Century COE Program
NAIST Graduate School of Information Science