Sangyun Lee

I’m an Assistant Professor in the Division of Future Convergence at Sungkonghoe University, Seoul, Republic of Korea. I currently have a focus on Artificial Intelligence, Computer Vision, and Robotics, encompassing everything from theoretical foundations to diverse industrial applications. My passion lies in exploring how these cutting-edge technologies can be integrated to innovate and advance our understanding and capabilities in the field.

Before accepting this position, I worked in the Samsung Electronics Co., Ltd. as a Senior Researcher. I studied for my Ph.D. at Yonsei University under the supervision of Euntai Kim and completed this in 2018 with a dissertation entitled ‘Deep Neural Network based Discriminative Feature Learning and Its Application to Multiple Object Tracking’.

Apart from my academic pursuits, I enjoy running, playing League of Legends, and exploring the world of wines.

news

Sep 26, 2025	’StoneGAT: A Robust Fall Detection Framework via Skeleton-Aware Feature Learning on Graph Attention Networks’ was accepted for publication in International Journal of Control, Automation and Systems!
Sep 1, 2025	A new project titled ‘A Study on Vision-Language-based Multi-Object Tracking’ started in Sep 2025. This work is supported by the Sungkonghoe University Research Grant of 2025.
Aug 1, 2025	’StoneGAT: Skeleton-aware Graph Attention Networks for Robust Fall Detection from Single-frame Poses’ was accepted for presentation in ICCAS 2025, Incheon, Korea!
May 19, 2025	’Feature Refinement With Vision State Space Modules for Tiny Object Detection’ was accepted for publication in Journal of Institute of Control, Robotics and Systems!
Dec 17, 2024	’DeepCampus: A Campus Tour Mobile Application Using Deep Learning-Based Vision Technologies’ was accepted for publication in The Transactions of the Korean Institute of Electrical Engineers!
Nov 11, 2024	’Making MambaVision Temporal: Leveraging TSM for Efficient Video Understanding’ was accepted for presentation in ICEIC 2025, Osaka, Japan!
Aug 1, 2024	A new project titled ‘Making Action Detection Practical: Leveraging 2D CNNs for Real-Time and Online Applications’ started in Aug 2024. This work is supported by the Sungkonghoe University Research Grant of 2024.
Jul 1, 2024	A new project titled ‘Real-time monitoring technology for identifying hazardous situations for workers in industrial settings using video-based AI’ started in Jul 2024. This work is supported by the Technology development Program funded by the Ministry of SMEs and Startups(MSS, Korea).
Dec 12, 2023	’Making TSM Better: Preserving Foundational Philosophy for Efficient Action Recognition’ was accepted for publication in ICT Express!
May 1, 2023	A new project titled ‘Development of Core Technology for Preventing Identity Theft’ started in May 2023. This work is supported by the SECUCHART GLOBAL Co., Ltd.
Apr 14, 2023	’D-TSM: Discriminative Temporal Shift Module for Action Recognition’ was accepted for presentation in UR 2023, Hawaii, USA!
Oct 17, 2022	’Extended Siamese Convolutional Neural Networks for Discriminative Feature Learning’ was accepted for publication in International Journal of Fuzzy Logic and Intelligent Systems!
Aug 1, 2022	A new project titled ‘A Context-aware Multiple Object Tracking with Temporal Shifted Transformer’ started in August 2022. This work is supported by the Sungkonghoe University Research Grant of 2022.
Mar 1, 2022	A new project titled ‘Research on AI integration technology for comprehensive context-aware multiple object tracking’ started in March 2022. This work is supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIT).

selected publications

StoneGAT: A Robust Fall Detection Framework via Skeleton-Aware Feature Learning on Graph Attention Networks

Soeun Chun, Seokjun Song, Doyeop Lee, and Sangyun Lee

International Journal of Control, Automation, and Systems, 2025
Feature Refinement With Vision State Space Modules for Tiny Object Detection

Sangyun Lee

Journal of Institute of Control, Robotics and Systems, 2025

Abs Bib HTML

Tiny object detection (TOD) plays a vital role in robotics and control systems, particularly in applications such as dronebased surveillance, autonomous navigation, and intelligent monitoring. Detecting extremely small objects in aerial imagery poses significant challenges due to limited spatial detail and insufficient contextual information. While feature pyramid networks (FPNs) have proven effective in handling multi-scale features, conventional FPN-based approaches often suffer from reduced spatial precision and semantic richness, factors critical for reliable tiny object detection. To address these limitations, we propose an enhanced FPN architecture that integrates deformable convolution (DeformConv) and vision state space modules (VSSM). DeformConv replaces standard fixed-kernel convolutions in the lateral connections, enabling adaptive focus on sparse spatial locations to better capture geometric variations and fine-grained spatial features. In parallel, VSSM modules are incorporated into the top-down pathway to model long-range contextual dependencies, thereby mitigating the semantic information loss typically encountered in traditional FPNs. Experiments conducted on the VisDrone2019 and AI-TOD v2 datasets demonstrate that the proposed architecture consistently outperforms conventional FPN-based detectors, validating the robustness and effectiveness of our feature refinement strategy for tiny object detection.
@article{vssmfpn, title = {Feature Refinement With Vision State Space Modules for Tiny Object Detection}, journal = {Journal of Institute of Control, Robotics and Systems}, volume = {31}, number = {7}, pages = {797-801}, year = {2025}, issn = {1976-5622}, doi = {https://doi.org/10.5302/J.ICROS.2025.25.0119}, author = {Lee, Sangyun}, keywords = {tiny object detection, feature pyramid network, vision state space model, mamba}, }
DeepCampus: A Campus Tour Mobile Application Using Deep Learning-Based Vision Technologies

Mujae Park, Yun A Kim, Heeju Cha, and Sangyun Lee

The Transactions of the Korean Institute of Electrical Engineers, 2025

Abs Bib HTML

This study proposes a campus tour mobile application called DeepCampus that leverages deep learning-based vision technologies to enhance user engagement and interaction. The proposed application introduces two core features: a photo mission, where users take photos at designated campus locations, and a commemorative photo creation function that transforms these photos into character-based styles. The photo mission relies on a CNN-based place recognition model, trained on a custom-built campus dataset, to verify mission completion. The commemorative photo creation function uses a diffusion model to generate personalized, stylized photos, providing users with unique records of their campus experience. By integrating place recognition and generative models, the application encourages active exploration. The effectiveness of the proposed system is validated through experiments on the custom dataset, demonstrating its technical reliability and potential to enhance campus tours.
@article{deepcampus, title = {DeepCampus: A Campus Tour Mobile Application Using Deep Learning-Based Vision Technologies}, journal = {The Transactions of the Korean Institute of Electrical Engineers}, volume = {74}, number = {1}, pages = {127-134}, year = {2025}, issn = {1975-8359}, doi = {https://doi.org/10.5370/KIEE.2025.74.1.127}, author = {Park, Mujae and Kim, Yun A and Cha, Heeju and Lee, Sangyun}, keywords = {Campus tour, Mobile application, Place recognition, Generative models, Commemorative photo}, }
Making TSM better: Preserving foundational philosophy for efficient action recognition

Seok Ryu, Sungjun Hong, and Sangyun Lee

ICT Express, 2024

Abs Bib HTML

In this study, we present the Discriminative Temporal Shift Module (D-TSM), an enhancement of the Temporal Shift Module (TSM) for action recognition. TSM has limitations in capturing intricate temporal dynamics due to its simplistic feature shifting. D-TSM addresses this by introducing a subtraction operation before the shifting. This enables the extraction of discriminative features between adjacent frames, thereby allowing for effective action recognition where subtle motions serve as crucial cues. It preserves TSM’s foundational philosophy, prioritizing minimal computational overhead and no additional parameters. Our experiments demonstrate that D-TSM significantly improves performance of TSM and outperforms other leading 2D CNN-based methods.
@article{RYU2024570, title = {Making TSM better: Preserving foundational philosophy for efficient action recognition}, journal = {ICT Express}, volume = {10}, number = {3}, pages = {570-575}, year = {2024}, issn = {2405-9595}, doi = {https://doi.org/10.1016/j.icte.2023.12.004}, url = {https://www.sciencedirect.com/science/article/pii/S2405959523001625}, author = {Ryu, Seok and Hong, Sungjun and Lee, Sangyun}, keywords = {Action recognition, Gesture recognition, Temporal modeling, Temporal shift module}, }
Multiple Object Tracking via Feature Pyramid Siamese Networks

Sangyun Lee, and Euntai Kim

IEEE Access, 2019

Abs Bib HTML

When multiple object tracking (MOT) based on the tracking-by-detection paradigm is implemented, the similarity metric between the current detections and existing tracks plays an essential role. Most of the MOT schemes based on a deep neural network learn the similarity metric using a Siamese architecture, but the plain Siamese architecture might not be enough owing to its structural simplicity and lack of motion information. This paper aims to propose a new MOT scheme to overcome the existing problems in the conventional MOTs. Feature pyramid Siamese network (FPSN) is proposed to address the structural simplicity. The FPSN is inspired by a feature pyramid network (FPN) and it extends the Siamese network by applying FPN to the plain Siamese architecture and by developing a new multi-level discriminative feature. A spatiotemporal motion feature is added to the FPSN to overcome the lack of motion information and to enhance the performance in MOT. Thus, FPSN-MOT considers not only the appearance feature but also motion information. Finally, FPSN-MOT is applied to the public MOT challenge benchmark problems and its performance is compared to that of the other state-of-the-art MOT methods.
@article{8587153, author = {Lee, Sangyun and Kim, Euntai}, journal = {IEEE Access}, title = {Multiple Object Tracking via Feature Pyramid Siamese Networks}, year = {2019}, volume = {7}, number = {}, pages = {8181-8194}, keywords = {}, doi = {10.1109/ACCESS.2018.2889442}, issn = {2169-3536}, month = {}, }