[English/Japanese]

  Takahiro OKABE

Ph.D. in Information Science and Technology
Department of Artificial Intelligence
Kyushu Institute of Technology
Professor



Contact
Address: 680-4 Kawazu, Iizuka, Fukuoka 820-8502, Japan
Office: E714 Departmental Research Building
Phone: +81-948297629
E-mail: okabe@ai.kyutech.ac.jp


Publications

[Recent Papers]

    Light Transport:
We propose a method for acquiring the multispectral LT (Light Transport) by using a single off-the-shelf multi-primary DLP projector; it does not require any self-built equipment, geometric registration, and temporal synchronization. Specifically, based on the rapid color switch due to a rotating color wheel in the projector, we present a method for estimating the spectral properties of the projector in a non-destructive manner, and a method for acquiring the images of a scene illuminated only by one of the primary colors.

    Diffuse-Specular Separation:
We propose a robust method for separating reflection components in a set of images of an object taken under multispectral and multidirectional light sources. We consider the set of images as the 3D data whose axes are the pixel, the light source color, and the light source direction, and then show the inherent structures of the 3D data. Based on those structures, our proposed method separates diffuse and specular reflection components by combining sparse NMF and SVD with missing data. We show that our method works better than some of the state-of-the-art techniques.

    Light Field Acquisition:
We propose an approach to measuring the 4D light field of a self-luminous extended light source by using an LC panel, i.e. a programmable filter and a diffuse-reflection board. Our proposed approach recovers the 4D light field from the images of the board illuminated by the light radiated from an extended light source and passing through the LC panel. We make use of the feature that the transmittance of the LC panel can be controlled both spatially and temporally, and recover 4D light fields efficiently and densely on the basis of multiplexed sensing and adaptive sensing.

    Image-Based Lighting:
We acquire the omnidirectional lighting environment of a real scene by using a hyperspectral camera, which has several tens of bands in visible spectrum, and then use the acquired spectral lighting environment for image synthesis. In particular, we compare the images synthesized by using the spectral lighting environment with those using the RGB lighting environment, and evaluate the difference between them. We demonstrate that the spectral lighting environment is highly important for rendering fluorescent materials.

    Super-Resolution:
We propose a MAP-based multiframe super-resolution method for flickering objects such as LED electronic message boards. Since they often flicker at low refresh rates, missing areas where LEDs are off during the exposure time of a camera by chance are observed. To suppress unexpected artifacts due to those missing areas, our proposed method detects outlier pixels on the basis of the spatio-temporal analysis of pixel values, and removes them from the MAP estimation by incorporating the weights of pixels into the likelihood term.

    Inverse Lighting:
We extend inverse lighting by taking an unknown and nonlinear radiometric response function of a camera into consideration, and propose a method for simultaneously recovering the lighting environment of a scene and the response function from a single image of an object. Through a number of experiments, we demonstrate that the performance of our proposed method depends on the lighting distribution, response function, and surface albedo, and address the conditions under which the simultaneous recovery works well.

    Shape and Reflectance Recovery:
We propose a method for simultaneously estimating the spectral reflectance and normal per pixel from a small number of images taken under multispectral and multidirectional light sources by integrating multispectral imaging and photometric stereo. In addition, taking attached shadows observed on curved surfaces into consideration, we derive the minimum number of images required for the simultaneous estimation and propose a method for selecting the optimal light sources in terms of noise.

[Publication List]



Profile

[Education & Work Experience]
2017.03 - Department of Artificial Intelligence, Kyushu Institute of Technology: Professor
2013.04 - 2017.02 Department of Artificial Intelligence, Kyushu Institute of Technology: Associate Professor
2012.11 - 2013.03 Institute of Industrial Science, The University of Tokyo (Sato Group): Project Associate Professor
2007.04 - 2012.10 Institute of Industrial Science, The University of Tokyo (Sato Group): Research Associate
2007.01 - 2007.03 Institute of Industrial Science, The University of Tokyo (Sato Group): Research Assistant
2001.01 - 2006.12 Institute of Industrial Science, The University of Tokyo (Sato Group): Technical Associate
2000.12Graduate School of Science, The University of Tokyo (Theoretical Astrophysics Group): Withdrawal from Ph.D. Program
1999.03Graduate School of Science, The University of Tokyo (Theoretical Astrophysics Group): Master of Science
1997.03Department of Physics, School of Science, The University of Tokyo: Bachelor of Science
[Others]
2017.04 - National Institute of Informatics: Visiting Professor
2013.04 - Institute of Industrial Science, The University of Tokyo: Cooperative Research Fellow
2014.04 - 2017.03 National Institute of Informatics: Visiting Associate Professor
2011.10 - 2012.09 Eberhard Karls Universität Tübingen (Computer Graphics Group: Prof. Hendrik Lensch): Visiting Scholar
2011.09Universität Ulm (Computer Graphics Group: Prof. Hendrik Lensch): Visiting Scholar
2011.03The University of Tokyo: Ph.D. in Information Science and Technology


Awards


Activities in Academic Societies

[Journal] [Conference etc.]

Last updated: November 13, 2017