About Me
My name is Seito, and I study computer vision & machine learning. I am currently a graduate student working on vision & language at Aoki Media Sensing Lab. in Keio University, and live in Tokyo, Japan. My main interest in the tech industry is creating software to give an impact on many people’s lives and experience.
Currently I am looking for engineering/research internship/full-time offers, especially ones in which I can express and enhance my skills. If you’re interested, please don’t hesitate to contact me!
Programming Languages: Python, C/C++, R, MATLAB, Java
Skills: Machine Learning, Computer Vision, Natural Language Processing, Data Science
Work Experience
Startup focusing on optimizing land usage using geospatial data
- Automated satellite data download and processing using Python, Gdal, GIS, Docker, and AWS
- Built a web application for visualizing geospatial data
- DevOps/MLOps, building infrastructure for application of machine learning
National Institute of Advanced Industrial Science and Technology (AIST)
https://www.aist.go.jp/index_en.htmlPart-time Research Assistant / Intern, AI Research Centre - Computer Vision Research Team
Apr 2019 - present
Japanese resesarch facility aiming to utilize scientific knowledge and technology to meet socio-economic needs
- Conducting research on the latest computer vision
- Publications in video recognition, vision & language, and cross-modal representation learning / retrieval
Amazon Web Services Japan
https://aws.amazon.com/Intern, Machine Learning Solutions Architect
Jan 2020
- Worked as a team and analyzed real-world problems seen through data
- Used Amazon S3, Amazon Sagemaker to conduct analysis and deploy scalable machine learning models
Publications
International Conferences
- S. Kasai, Y. Ishikawa, M. Hayashi, Y. Aoki, K. Hara, H. Kataoka, Retrieving and Highlighting Action with Spatiotemporal Reference, IEEE International Conference on Image Processing (ICIP), Oct 2020.
- N. Akimoto, S. Kasai, M. Hayashi, Y. Aoki, 360-Degree Image Completion by Two-Stage Conditional GANs, IEEE International Conference on Image Processing (ICIP), Sep 2019.
- K. Yoshida, K. Okayasu, K. Tsubura, M. Minoguchi, T. Ikeya, A. Nakamura, S. Kasai, H. Kataoka, Which Generates Better Jokes, Hand-Crafted Features or Deep Features, CVPR 2019 Language & Vision Workshop, Jun 2019.
Domestic Journals
Domestic Conferences
- 笠井誠斗，林昌希，秋本直郁，青木義満，自然言語指示文を用いた顔画像属性変換，動的画像処理実利用化ワークショップ (DIA), OS5-3，2019年3月
- 笠井誠斗，中込優，黒澤郁音，矢代達希，西川真史，相澤宏旭，福井宏，【招待ショートサーベイ】コンピュータビジョンと自然言語処理の共存マップ， 電子情報通信学会，パターン認識・メディア理解研究会(PRMU)，2018年10月
- S. Kasai, Y. Ishikawa, T. Wakamiya, K. Hara, H. Kataoka, Exploring the Best Model for Dense Captioning Events in Videos, The 22nd Meeting on Image Recognition and Understanding (MIRU), Aug 2019.
Education
Keio University, Faculty of Science and Technology
https://www.st.keio.ac.jp/en/BS in Electronic Engineering
2015 - 2019
Courses taken include Image Processing, Neural Networks, Electromagnetic Physics, Information Theory, Electronic circuits, and Semiconductors. My graduate thesis theme is based on vision and language. Outside of my studying hours, I was captain in the tennis club, which enhanced my managing and problem solving skills as well as my leadership.
- Advisor: Prof. Yoshimitsu Aoki
- Lab homepage: Aoki Media Sensing Lab.
- Student Interview (Japanese)
Graduate School in Keio University, Faculty of Science and Technology
https://www.st.keio.ac.jp/en/MS in Computer Science
2019 - present
Mainly taking courses related to computer science, such as compilers, networks, algorithms and data structures, software engineering.
- Advisor: Prof. Yoshimitsu Aoki
- Lab homepage: Aoki Media Sensing Lab.
Projects
This organization consists of members from various universities in Japan all studying computer vision. Over 100 members participate in the projects that we hold at cvpaper.challenge, e.g. CVPR 2018 survey.
- Participating as a research member for publication in top conferences on computer vision and natural language processing
- Creating survey slides and material for computer vision
ActivityNet Large-Scale Activity Recognition Challenge (Kinetics, Dense-Captioning Events in Videos)
http://activity-net.org/challenges/2019/challenge.htmlApr 2019 - Jun 2019
Most natural videos contain numerous events. For example, in a video of a 'man playing a piano', the video might also contain another 'man dancing' or 'a crowd clapping'. This challenge studies the task of dense-captioning events, which involves both detecting and describing events in a video.
- Participated in the competition @CVPR2019 WS
- 9th in video classification task
A lecture given on AI programming using PyTorch. The audience was students from Keio University MBA.
- Conducted the whole lecture, including making the code and preparing the slides and presentation
- Explained the powers and limits of AI to beginners without giving high-level mathematics
- Used Python, PyTorch to build a classification algorithm with visual reasoning
A Little More About Me
I used to live in Miami, FL in the U.S. for about 6 years when I was young, so I’m comfortable communicating in both Japanese and English. Alongside my interests in machine learning research and software engineering some of my other interests and hobbies are:
- Playing Tennis (I’ve played for more than 17 years!)
- Travelling