Biography

I am currently a Ph.D. candidate with the Department of mechanical and information engineering and the Department of Smart Cities, University of Seoul, Korea. My research interests include on-device deep learning, prognostics and health management, and application-specific parallel computing.

Skills

• Deep learning and machine learning (PyTorch, Scikit-Learn)

• Embedded system (sensor and motor control, on-device AI)

• Digital signal processing

• Parallel computing (OpenMP, NVIDIA CUDA Toolkit)

• Academic and technical writing in English

• Programming languages: C/C++, CUDA C++, Python, R, MATLAB

Publications

International Journal Article

1. Seongjae Lee and Taehyoun Kim, “FRFconv-TDSNet: Lightweight, Noise-Robust Convolutional Neural Network Leveraging Full-Receptive-Field Convolution and Time Domain Statistics for Intelligent Machine Fault Diagnosis,” IEEE Transactions on Instrumentation and Measurement, vol. 73, pp. 1–13, 2024. [DOI]

2. Seongjae Lee, Taewan Kim, and Taehyoun Kim, “Multi-domain Vibration Dataset with Various Bearing Types Under Compound Machine Fault Scenarios,” Data in Brief, Accepted (in press), 2024.

3. Seongjae Lee and Taehyoun Kim, "Impact of Deep Learning Optimizers and Hyperparameter Tuning on the Performance of Bearing Fault Diagnosis," IEEE Access, vol. 11, pp. 55046–55070, 2023. [Source Code] [DOI]

4. Seongjae Lee and Taehyoun Kim, "Parallel Dislocation Model Implementation for Earthquake Source Parameter Estimation on Multi-Threaded GPU," Applied Sciences, vol. 11, no. 20, p. 9434. 2021. [DOI]

5. Seongjae Lee and Taehyoun Kim, “Search Space Reduction for Determination of Earthquake Source Parameters Using PCA and k-Means Clustering,” Journal of Sensors, vol. 2020, pp. 1–12, 2020. [DOI]

6. Jinho Park, Taeyoung Yoo, Seongjae Lee, and Taehyoun Kim, "Urban Noise Analysis and Emergency Detection System using Lightweight End-to-End Convolutional Neural Network," International Journal of Computers Communications & Control, vol. 18, no. 5, p. 5814, 2023. [DOI]

7. Taeyoung Yoo, Seongjae Lee, and Taehyoun Kim, "Dual Image-Based CNN Ensemble Model for Waste Classification in Reverse Vending Machine," Applied Sciences, vol. 11, no. 22, p. 11051, 2021. [DOI] 

Domestic Journal Article

8. Youngjun Kim, Taewan Kim, Suhyun Kim, Seongjae Lee, and Taehyoun Kim, "Lightweight On-Device AI-Based Fault Diagnosis System Using Continual Learning," IEMEK Journal of Embedded Systems and Applications, vol. vol.19, no.3, pp. 151–158, 2024.

9. Minha Lee, Seongjae Lee, and Taehyoun Kim, "Performance Evaluation of Efficient Vision Transformers on Embedded Edge Platforms," IEMEK Journal of Embedded Systems and Applications, vol. 18, no.3, pp. 89–100, 2023.

Domestic Conference

10. Youngjun Kim, Taewan Kim, Suhyun Kim, Seongjae Lee, and Taehyoun Kim, "On-Device Machine Fault Diagnosis System Using Continual Learning," Institute of Embedded Engineering of Korea Autumn Conference, pp. 56-60, Nov, 2023.

11. Seongjae Lee, Heetae Lee, and Hyunggu Jung, "Link-it!: Designing an App Managing Website Links Through a Design Thinking Process," The HCI Society of Korea, The Proceedings of HCI KOREA 2023, pp. 654–659 Feb, 2023.

12. Jaeyoung An, Taeyoung Yoo, Jihwan Yoo, Seongjae Lee, Taehyoun Kim, and Sooil Lee, "Implementation of on-site mobile vibration measurement system using digital accelerometer," The Korean Society for Noise and Vibration Engineering Autumn Conference 2020, p. 269, 2020.

Projects

On-Device AI

1. Design and Implementation of Edge-Based Real-Time Adaptive Machine Fault Monitoring System (May 2024–Apr. 2025)

• • (Ongoing Project) Develop a real-time, on-device, and domain-adaptive machine fault monitoring system leveraging the on-device training technique.

2. Deep Learning-Collaborative Open-Source Intelligent Fault Diagnosis System based on Open IoT Platform and Edge Computing Environment (Mar. 2022–Feb. 2025)

• Develop a vibration-based intelligent fault diagnosis system, including dataset generation [2], model benchmarking platform in noisy environments [3],  lightweight and noise-robust model development [1], and continual learning processes [8].

• Perform additional on-device AI studies related to the project [6, 9].

3. AI-based Coffee Bean Automatic Sorting System (Aug. 2021–July 2022)

• Develop a deep learning-based automatic coffee bean sorting machine and its control system  (500 ms per single bean classification, 99% accuracy).

GPU Computing

1. R-MDPS Motion Visualization System Development (June 2023–Mar. 2024)

• • Implement a displacement tracking software based on a GPU using a high-resolution smartphone camera and the phase-based motion magnification method (max. error of 0.169 mm compared to LVDT sensor).

2. Performance Improvement of Remote Sensing Applications using CUDA GPU-based Parallelism (May 2020–Apr. 2021)

• • Develop a parallel earthquake dislocation model using the NVIDIA CUDA to accelerate a source parameter inversion process [4].

3. Development of Estimation Technique of Earthquake Model parameters Through Fusion of Radar and Seismic data (Jan. 2019–Dec. 2019)

• Implement a C++-based earthquake model parameter estimation software.

• Develop a machine learning-based search space selection algorithm that accelerates nonlinear earthquake parameter inversion [5].

Experiences

1. Instructor, AICOSS deep learning bootcamp, University of Seoul (2022, 2023)

2. Teaching Assistant, Embedded Systems and Experiment, University of Seoul (2020–2024)

3. Teaching Assistant, Programming Methodolody and Practice, University of Seoul (2021–2024)

4. Teaching Assistant, Digital Logic Design and Experiments, University of Seoul (2018–2020)

Open Source Projects

Fault Diagnosis Optimizer Benchmark

Educations

University of Seoul / Integrated Masters and Doctoral Program

Sep. 2019–Present, Seoul, Korea (expected graduation: Aug. 2025)

University of Seoul / Bachelor of Engineering (Department of Mechanical and Information Engineering)

Mar. 2015–Aug. 2019, Seoul, Korea

Contact Me