Interest
- Computer Vision
- Deep Learning
- Web Development
Education
- Ph.D. Student in Computer Science2023.03 - current Computer Science Program, Graduate School of Science and Technology, Meiji University Image Processing Laboratory (IPLab)
- Master's Degree in Computer Science2021.04 - 2023.03 Computer Science Program, Graduate School of Science and Technology, Meiji University Image Processing Laboratory (IPLab)
- Bachelor's Degree in Computer Science2017.04 - 2021.03 Department of Computer Science, School of Science and Technology, Meiji University Image Processing Laboratory (IPLab)
Work Experience
- Qdoga Inc.
- Meiji University2021.04 - 2023.03 Teaching Assistant
- SensAI
- Yahoo Japan 2022 Summer Internship
- RECRUIT Job for Student 2022 Autumn ~Engineer/Data Specialist~
- SPIC
- Meiji University2023.04 - current Research Associate
Publication
- A study on accuracy improvement of small object detection using deep neural networksIEEE Asia Pacific Conference on Circuit and Systems, APCCAS 2021, Penang, Malaysia, November 22-26, 2021
- Testing environment for developing a wireless networking system based on image-assisted routing for sports applicationsProceedings of the 9th International Conference on Sport Sciences Research and Technology Support, icSPORTS 2021, October 28-29, 2021
- 深層学習に基づく小物体検出の精度向上に関する一検討電子情報通信学会(IEICE) スマートインフォメディアシステム研究会（SIS), 2022.12
- 意味論的領域分割の計算コストと精度の関係に関する研究電子情報通信学会(IEICE) スマートインフォメディアシステム研究会（SIS), 2022.12
- 空撮画像を対象としたInformed-Filtersに基づく人検出のC++AMPを用いたリアルタイム実装画像電子学会 画像電子学誌 第52巻 第2号 pp.329-338, 2023.4
- Human Pose Estimation Based on Part-Wise 6-DoF Pose Estimation2023 International Workshop on Smart Info-Media Systems in Asia, pp.77-81, 2023.9
- Dataset Creation for Semantic Segmentation Using Colored Point Clouds Considering Shadows on Traversable AreaJournal of Robotics and Mechatronics, 2023.12
- Data Augmentation for Semantic Segmentation Using a Real Image Dataset Captured around the Tsukuba City HallJournal of Robotics and Mechatronics, 2023.12
- 遠隔協調作業における臨場感向上のための単眼カメラによる部屋の3次元構造の推定ライフインテリジェンスとオフィス情報システム研究会(LOIS), 2024.03
- 3D Structure Estimation of Room Environment Using Semantic Segmentation and Model FittingIEEE International Workshop on Metrology for Living Environment (MetroLivEnv) , 2024.06
- 合成画像を訓練サンプルとする分類器のバイラテラルフィルタによる精度向上電子情報通信学会(IEICE) スマートインフォメディアシステム研究会（SIS), 2024.10
- Improving Resolution of Translated Infrared ImagesIEEE SENSORS, 2024.10
- Creation of Day-Night Visible-Thermal Paired Image Dataset via Image RegistrationIEEE IRC, 2024.12
- AMD Ryzen CPUを対象としたCNNにおける積和演算の並列実装第50回 パルテノン研究会, 2024.12
- CNN における積和演算の高速化に関する一検討第50回 パルテノン研究会, 2024.12
- バーチャル内見を目的とする3次元構造推定のための効率的なセマンティックエッジ抽出電子情報通信学会(IEICE) MVE研究会（SIS), 2025.03
- 屋内環境における単一画像からの3次元情報推定手法の回転耐性の向上電子情報通信学会(IEICE) ライフインテリジェンスとオフィス情報システム研究会(LOIS), 2025.03
Award
- グッドプレゼンテーション賞「遠隔協調作業における臨場感向上のための単眼カメラによる部屋の3次元構造の推定」ライフインテリジェンスとオフィス情報システム研究会(LOIS), 2024.03
- 優秀プレゼンテーション賞「AMD Ryzen CPUを対象としたCNNにおける積和演算の並列実装」第50回 パルテノン研究会, 2024.12
Skills
- Python
- PyTorch
- TensorFlow
- FastAPI
- JavaScript
- React
- Svelte
- Astro
- Docker
- AWS
- Swift
- SwiftUI
- Ruby
- Sinatra
- PHP
- CodeIgniter
- Blender
- ONNX
- TensorRT
- C++
Projects
Small Object Detection using Deep Learning
Various methods for object detection based on deep learning have been proposed, but there is still much room for improvement in the accuracy of small object detection. This research aims to improve the detection accuracy of small objects by examining what processing is important for improving accuracy. Based on the results of this examination, we propose a new method to improve the accuracy of small object detection.
3D Room Structure Reconstruction using Monocular Camera
In recent years, with the spread of VR devices and technological improvements, the demand for technology that reflects real space in virtual space has increased. In particular, technology that reflects indoor environments in virtual space is used in various fields such as architecture, interior design, and robotics. To reflect real space in virtual space, it is necessary to estimate the 3D structure of real space, but expensive equipment is required, and there are problems in terms of computational complexity and ease of measurement. Therefore, in this research, we limit the estimation target to indoor environments and propose a 3D structure estimation method using a single image.6DoF Pose Estimation using Deep Learning Models
This research aims to devise a high-precision 3D object position estimation method. High-precision 3D object position estimation has various challenges such as datasets for training, handling occlusions, and camera calibration. This research aims to achieve everything from dataset creation to practical application of models.
Video Streaming Service Development
Developed a video streaming service for iOS/Android. Used Ruby's Sinatra for the server side and AWS for infrastructure. Used Firebase Authentication for authentication. I was responsible for backend API development, Firebase Authentication implementation, iOS app implementation with SwiftUI, and CI/CD construction using Xcode Cloud and GitHub. In particular, I implemented the entire iOS app migration from UIKit to SwiftUI from scratch by myself.Search System Development using NLP Embedding Models
Developing a search system using embedding models with BERT and Faiss used in natural language processing. Used FastAPI for the server side and both AWS and on-premises for infrastructure. The search target data is 60TB in PDF format, and high-speed search is required for these documents. Creating HuggingFace Datasets, training tokenizers and models, and building custom models.
SPIC Inc. Internal System Development
Building internal systems for SPIC Inc., which develops and sells health-related products including Lypo-C, beauty-related products, and original brand products.Contact
mjun __at__ mjunya.com