[var] => content
[type] => preview
[var] => cut_url
Takanohara Ekinishi Danchi , Kabutodai , Kizugawashi, Kyoto, Japan, Objective Education
Tel: + Mobile: + Email:
A researcher position in the field of D/D Vision. PhD in Computer Vision Technical University of ClujNapoca, Romania Thesis title: D Modeling Techniques for UltraRealistic Systems. MSc in OptoMechatronics Wakayama University, Japan Thesis title: A Method for Reconstructing Structure from Omnidirectional View Sequence without Feature Matching. BSc in Electrical Engineering Technical University of ClujNapoca, Romania Thesis title: Acquisition of Data from Teletext Signal of a TV Broadcast.
Tackled issues of camera calibration as well as D/D map reconstruction from a multiview geometry setup of omnidirectional cameras. Investigated statistical optimization techniques in order to obtain a measurement of the reliability of essential matrices in a multiple view calibration scenario. Performed full D reconstruction of small objects from multiple range views captured from unknown locations. Investigated the maximum displayable resolution for two autostereoscopic D displays: gCubik and fVisiOn. Omni Directional Cameras: calibration, D/D reconstruction. Range Views: automatic registration, refining of the D structure. Genetic Algorithm: solving pathplanning problems for robots.
Computer programming: Object Oriented Programming years, C/C++ years, Matlab years, OpenGL years. Assembly Language. Computer vision: Familiar with OpenCV, VXL, SIFT, SBA package for refining of D structure; Developed SfM and SLAM applications for perspective and omnidirectional cameras; Proficient in Linear Algebra, Projective Geometry and Optimization. Development tools: Visual C++, C++ Builder. Operating Systems: Linux RedHat, Fedora Core, MS Windows. Hardware: Designed PCBs; Implemented drivers for data acquisition and robot control; Experience with IC communication protocol and programmable devices microcontrollers, display drivers. Languages: English fluent, Japanese conversational: JLPT Level , German basic.
Researcher at National Institute of Information and Communications Technology NICT Keihanna Lab., Japan
Investigated the maximum displayable angular resolutions in the horizontal and vertical directions of a planar D object displayed by a glassesfree tabletop D display that is using a conicshaped optical device and an array of projectors.
Involved in a project that performs automatic reconstruction of full D models of real objects from range views. The application aligns unregistered D views taken from unknown locations.
Unregistered D views D images
Image Based Registration
Invariant D features extracted by SIFT
Full D model
Researcher at Wakayama University, Japan
Performed D reconstruction of an unknown environment by using an omnidirectional camera. The process involved generating a virtual camera with a cylindrical image plane.
Original omni directional image
Virtual omni directional camera
Calibration of the virtual omnidirectional
Matching color histograms
Rough D reconstruction of the depth to the surrounding objects
Subsequently, color histograms of projections from two optical centers on planar screens were matched. By incrementing the distance and by varying the slanting angle, the two projections would match on the surface of the object.
Journal Publications Patent Conference Publications selection
Daniel M., Takahiro Miyashita and Hiroshi Ishiguro, A Method for Reconstructing Structure from Omnidirectional View Sequence without Feature Matching in IPSJ Journal of CVIM, July, . Daniel M., Sumio Yano, Yoshida Shunsuke, Naomi Inoue, Method and apparatus for registering D range data, submitted to Japan Patent Office, Registration Number: . Daniel M., Shunsuke Yoshida, Masahiro Kawakita and Hiroshi Ando, Analysis of resolution limitation of glassesfree D tabletop display, IS&T/SPIE Electronic Imaging , Stereoscopic Displays and Applications XXII Conference , San Francisco, U.S.A, January . Daniel M., Roberto LopezGulliver, Sumio Yano and Hiroshi Ando, Online System for D Model Reconstruction and Visualization on gCubik AutoStereoscopic Display, Proceedings of DSA, Tokyo, Japan, May thst, . Daniel M., Sumio Yano, Naomi Inoue, Automatic Registration of Multiple Range Views based on Feature Matching and FeatureDepth Correspondences, Proceedings of ICCP , ClujNapoca, Romania, August thth, . Jaume VergesLlahi, Daniel M., Toshikazu Wada, A New Reliability Measure for Essential Matrices Suitable in Multiple View Calibration, Proceedings of VISAPP , nd th of January , vol., , Funchal, Portugal. Daniel M. and Toshikazu Wada, A calibrated pinhole camera model for single Viewpoint Omnidirectional Imaging Systems, Proceedings of ICIP , October , Singapore. H. Yamamoto and D. M., Robot Path Planning for Visiting FA Working Points by Obstacle Avoiding using GA, Artificial Life and Robotics , Tokyo, Japan, January, ,
Marital Status: MARRIED, CHILD.