AtCoder: CyanUse on a Daily Basis
- Japanese (Native)
- Python
- C / C++
- Vitis / Vivado HLS
- JavaScript
- bash / zsh
- 日本語
- English
I'm Nobuho Hashimoto,
a graduate student
in the University of Tokyo
(Takamaeda Lab).
I'm researching on acceleration of
image and point-cloud processing on an FPGA.
(Natural / Programming) Languages
These are the languages I am using and have been used so far. But I am eager to try other languages and can become familiar with them soon based on my experience.
Use Sometimes
- English (Business)
- HTML / CSS
- PHP
- System Verilog
- OCaml
Have Used Before
- Spanish
- Thai
- Chinese
- MySQL / PostgreSQL
- Ruby
- Java
- Scheme
- VHDL
- Prolog
Favorites
Programming
Travel
Soccer
Ski / Snowboard
Japanese Sake
Wine
Diving
Swimming
Fishing
| 2022/06 | PADI Open Water Diver (OWD) Certification Card |
| 2022/06 | ANSA Wine Examination Grade 3 |
| 2022/03 | SSI Japanese Sake Examination Grade 3 |
Research / Education
Research
In the graduation thesis, I worked on accelerating an image processing filter called the bilateral filter using reconfigurable hardware (FPGA). Although image processing technology is applied in many fields such as medical diagnosis and self-driving, it is still difficult to process high-resolution images fast and small-scale. Therefore, I proposed its accelerator using parallelization and pipeline techniques, implemented on an FPGA, and verified that it was smaller and could process faster than the other existing approaches.
In the image processing fields, ML (Machine Learning) technology has been widely incorporated recently. Still, the technology requires more computation and execution time than the existing technologies. Therefore, I aim to create a helpful technology by building an acceleration mechanism specific to ML-based image and point-cloud processing technologies.
Presentations
| 2022/09 |
“FPGAを用いたフルパイプラインによるバイラテラルフィルタの高速化手法” @ Forum on Information Technology (FIT) (Information Processing Society of Japan (IPSJ), etc.) Program and Abstract |
| 2022/03 |
“機械学習ベースの動画像処理における近似計算手法の検討” @ Technical Committee on Computer Systems (CPSY) (The Institute of Electronics, Information and Communication Engineers (IEICE)) Program and Abstract, Paper (Copyright ©2022 by IEICE), Slides (below) (only in Japanese) |
| 2021/09 |
“An FPGA-Based Fully Pipelined Bilateral Grid for Real-Time Image Denoising” (Full Paper) @ The International Conference on Field-Programmable Logic and Applications (FPL) (The top conference on FPGA) Program and Abstract, Paper (IEEE Xplore), Paper (arXiv), Slides (below), Movie |
| 2021/06 |
“FPGAを用いたフルパイプラインによるバイラテラルフィルタの高速化手法” @ Technical Committee on Reconfigurable Systems (RECONF) (The Institute of Electronics, Information and Communication Engineers (IEICE)) “Best Presentation Award (young division)” Program and Abstract, Paper (Copyright ©2021 by IEICE), Slides (below) (only in Japanese) |
Education
| 2021/04 - present | Department of Computer Science, Graduate School of Information Science and Technology, The University of Tokyo |
| 2019/04 - 2021/03 | Department of Information Science, School of Science, The University of Tokyo |
| 2017/04 - 2019/03 | Natural Sciences I, College of Arts and Sciences, The University of Tokyo |
| 2014/04 - 2017/03 | Seifu Nankai High School |
| 2014/07 - 2014/08 | Brisbane Grammar School ( |
Work
| 2021/04 - present | Research Assistant at the University of Tokyo |
| Research on a multi-paradigm HLS (High-Level Synthesis) framework, which is easy to optimize the performance. | |
| 2021/10 - 2022/03 | Teaching Assistant at the University of Tokyo |
| Assistance in the class of the Processor and Compiler Lab in the Department of Information Science. | |
| 2021/04 - 2021/07 | Teaching Assistant at the University of Tokyo |
| Assistance in the class of the Hardware Lab in the Department of Information Science. | |
| 2020/01 - 2020/08 | Long-term internship at A Co., Ltd. |
| Reverse Engineering of a product written in the C language. Development of an image processing AI using Python. Development of Websites and APIs using Django. | |
| 2019/08 | Summer internship at M Ltd. |
| Proposal of a brand-new Web service and its development in a group using GitHub and AWS (see here). (Win the first prize.) | |
| 2019/02 - 2019/03 | Long-term internship at BizWings (Thailand) Co., Ltd. in Thailand |
| Introduction of a business efficiency improving system. | |
| 2017/04 - 2019/03 | Project Leader at AIESEC in the University of Tokyo |
Operation of global internship programs as the incoming exchange department.
|
|
| 2018/03 - 2018/05 | Long-term internship at M Co., Ltd. |
| Blockchain survey writer. |
Portfolio
Replacer
Development of automatic string converter with multiple conditions using JavaScript. The demo and implementation details are available here.
Nested UNet for Building Detection
Development of the Nested UNet for Building Detection using PyTorch. Accuracy is improved by using Fully Connected CRF and denoisng. Reference: “UNet++: A Nested U-Net Architecture for Medical Image Segmentation” by Zongwei Zhou et al. (Springer International Publishing 2018)
Multiple Images to Single PDF
Development of a converter from multiple images to a single merged PDF file using Python.
4-stage pipelined RISC-V processor with privileged instructions
Development of 4-stage pipelined RISC-V processor with privileged instructions using System Verilog. Branch prediction using Two-level adaptive predictor and forwarding are also implemented to reduce stalls and accelerate the processor. Exception and interruption handlers are implemented by using CSR and privileged instructions (mret).
ChipNet for LiDAR Segmentation
Development of the ChipNet for road segmentation from LiDAR point-clouds using PyTorch. KITTI Road dataset is used. Reference: “ChipNet: Real-Time LiDAR Processing for Drivable Region Segmentation on an FPGA” by Yecheng Lyu et al. (TCAS1 2019)
ResNet
Development of the ResNet for object recognition using PyTorch. 96.01 % accuracy in CIFAR-10 dataset. Reference: “Deep Residual Learning for Image Recognition” by Kaiming He et al. (CVPR 2016)
Multilayer Perceptron
Development of Multilayer Perceptron (MLP) for handwriting recognition from scratch using Python. 98.78 % accuracy in MNIST dataset.
Machine Learning from scratch
Development of various machine learning methods, such as Regression and Classification from scratch using Python.
Gaussian Filter on an FPGA
Development of the Gaussian Filter on an FPGA using Vitis HLS, Vivado, and PYNQ.
Image Captioning AI
Development of an image captioning AI in PyTorch. Reference: “Show and Tell: A Neural Image Caption Generator” by Vinayls et al. (ICML 2015)
Edge Layout Game
Development of a user-friendly game considering UI (User Interface) in JavaScript and HTML / CSS. The Website is available here (only in Japanese).
CG Projects
Various implementations related to CG (Computer Graphics), including parametric curves, ray tracing, image filtering, and so on in JavaScript and HTML / CSS. The Website is available here.
Synner
Development of an automatic data generation support Websites in JavaScript and HTML / CSS. Reference: “Is this Real? Generating Synthetic Data that Looks Real” by Mannino et al. (UIST 2019) The Website is available here.
RISC-V CPU
Group development of a CPU on an FPGA based on RISC-V. Development of FPU (Floating Point Unit) in System Verilog. Development of FPU simulation in C.
Various lifehacks on a Raspberry Pi
Remote light controller using infrared LEDs. Voice operation using Julius. NAS (Network Attached Storage). Private Websites using Nginx and WordPress.
Tetris
Development of a Tetris game in JavaScript and HTML / CSS.
OCaml Interpreter
Development of an OCaml interpreter in OCaml.
Othello
Development of a strong Othello player in OCaml.
Group Allocator
Development of a complex group allocation system in Python.
Slides
“機械学習ベースの動画像処理における近似計算手法の検討 (CPSY 2022/03)” (only in Japanese)
“FPGAを用いたフルパイプラインによるバイラテラルフィルタの高速化手法 (RECONF 2021/06)” (only in Japanese)
“ブロックチェーンとはどのような技術か” (only in Japanese)
Achievements
| 2022/06 | 3rd grade Certified Skilled Professional of Financial Planning |
| 2021/06 | Best Presentation Award (young division) in Technical Committee on Reconfigurable Systems (RECONF) |
| 2021/03 | Education Network for Practical Information Technologies (enPiT) |
| 2020/12 | Applied Information Technology Engineer Examination |
| 2020/02 | TOEFL iBT Score: 98 |
| 2017/10 | Driver's License |
| 2017/08 | TOEIC Score: 910 |
| 2016/08 | Silver prize winner (only in Japanese) in Japan Physics Olympiad in 2016 |
| 2015/11 | The EIKEN Test in Practical English Proficiency Grade Pre-1 |
| 2014/07 | The Japan Kanji Aptitude Test Grade 2 |
Others
| 2021/05 - present | Accountant in the Miyabi-kai (only in Japanese) (Alumuni group of the Ito Scholarship Foundation (only in Japanese)). |
| 2021/04 - present | Scholarship graduate student in the Ito Scholarship Foundation. |
| 2017/04 - 2021/03 | Scholarship student in the Ito Scholarship Foundation. |
| 2019/04 - 2020/03 | Executive committee member to organize a two days workshop in the Ito Scholarship Foundation. |
| 2019/09 | Participation in “Scandinavian Blockchain Study Tour with Sota Watanabe” (only in Japanese) in Estonia and Finland. |
| 2019/08 | Instructor of science classes in “Japan-Australia Science Laboratory” (only in Japanese) in Australia. |
| 2019/05 | Keynote speech in “Symposium on Blockchain and Exchange Meetings” (only in Japanese). Slides are shown above (only in Japanese). |