LeafWindow Inc.

About LeafWindow Inc.

LeafWindow Inc. is a  self-employed software development company established on November 22, 2021.

  • CEO and Software Engineer: Takuya Fukagai
  • Address: 2-5-24, Iwasaki, Gifu-city, Gifu 502-0005, Japan
  • Phone: +81-58-231-5272
  • Mobile Phone: +81-90-7824-4110
  • FAX: +81-58-232-8246
  • E-Mail: fukagai@leafwindow.com

About Me

I have been developing in-house software.

  1. Windows application with C#, Xaml(WPF), Web API and SQL
  2. Web applications with php, html, css and JavaScript
 
 

As a software engineer, I have the following experiences.

  • Research on garbage collection (automatic memory management algorithm) 2003-2004
    • Evaluated the performance of basic garbage collection methods for Java Virtual Machine that is used by embedded devices (J2ME CDC).
    • Implemented reference counting garbage collection and compared its performance with other basic garbage collection methods.
    • Read the book, “Garbage Collection: Algorithms for Automatic Dynamic Memory Management” written by Richard Jones and Rafael Lins.
  • Research and Development of Java Card Virtual Machine 2004-2005
    • Improved the performance of Java Card Virtual Machine for a RISC processor.
    • Modified source code of Java Card Virtual Machine so that it can execute Java Card application more efficiently.
  • Design and Development of Java Card application development environment 2005-2006
    • Designed and developed Java Card application development environment as an eclipse plugin.
  • Research and development on Linux Kernel
  • Development of visual image processing application
  • Development of software modules for mobile robots 2009-2011
    • Developed reusable software components for mobile robots.
    • These components use probabilistic data analysis methods like Kalman filter and particle filter. These methods are used to estimate the position and direction of mobile robots.
  • Research on tracking algorithm 2011-2013
    • Studied algorithms for tracking multiple moving objects among image sequences. These images contain clutter and the sensor sometimes cannot observe the moving objects. To solve this problem, a classical probabilistic tracking algorithm is applied. The method was evaluated through simulation. I wrote two patents on this research.
    • Read more than 30 titles of research papers and several books related to object tracking algorithm. Explained the details of these algorithms to clients and prepared power point slides, which describe the key points of these algorithms.
  • Research and development on Cordova plugin for iOS 2013-2014
    • To provide JavaScript programmer with easier access to sensor related technologies, I tested Cordova plugin with iOS. I wrote codes with Objective-C and JavaScript.
  • Research on Multiple-Target-Tracking algorithm 2013-2014
  • Research and development on step counting algorithm for 3-axis accelerometer
  • Research on Deep Learning and hardware designs that is suitable for deep learning algorithms.
    • Research and development on Object Detection Neural Networks
    • Research on deep learning algorithms including efficient convolution algorithms
    • Research on efficient simulation of Maxwell’s equations with deep learning
    • Research on movie compression algorithms

 

  • Education
    • MS Computational Intelligence and Systems Science, Tokyo Institute of Technology, Tokyo, Japan 2000
    • BS Physics, Tokyo Institute of Technology, Tokyo, Japan 1998

 

  • Computer Skills
    • Programming Languages: C#, WPF(xaml), Java, C, C++, CUDA, Objective-C, Python, Matlab (Octave), JavaScript, Perl, SQL, R
    • Platforms: Linux, Mac OSX, iOS, Windows, Unix