Publications

Publications (2021–Present)

2026

• PCEval: A Benchmark for Evaluating Physical Computing Capabilities of Large Language Models.

Inpyo Song, Eunji Jeon, and Jangwon Lee.
In Advances in Neural Information Processing Systems (NeurIPS), Sep 2026. (Poster, 25.8% acceptance rate).

Paper  |  Project Page  |  arXiv  |  Code

• Age-differentiated Multimodal Style in Youtube Thumbnails and Titles: A Feature-level Analysis of Visual and Textual Signals.
  Heeseok Yang, Yena Lee, Jeongbin Lee, Jungeun Baik, and Jangwon Lee.
  In Association for Information Science & Technology (ASIS&T) Annual Meeting, Nov 2026. (Poster, Accepted).

• Bounding-box Trajectories Matter for Video Anomaly Detection. 

Inpyo Song and Jangwon Lee.
In European Conference on Computer Vision (ECCV), Sep 2026. (Poster, 25.7% acceptance rate).

Paper  |  Project Page  |  arXiv  |  Code 

• Dualsplat plus : Enhanced Streaming 3d Gaussian Splatting via Adaptive Foreground-Background Decomposition.
  Hyemin Hwang, Minho Yeom, Hannah Cho, and Jangwon Lee.  
  KSII Transactions on Internet and Information Systems (TIIS), Jun 2026.
  Paper  |  Project Page  |  arXiv  |  Code 

• Human Pose Estimation in Videos using the Concept of a PID Controller.
  Yeeun Kim, Jongmin Lee, and Jangwon Lee.
  In Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD), Jun 2026. (Oral, 25.2% acceptance rate).
  Paper  |  Project Page  |  arXiv  |  Code 

2025

• Mindfulbreath: Every Breath You Take, Every Mind You Wake.
  Hannah Cho, Jingyeong Park, Yeonjun Seong, Hayoung Oh, Jangwon Lee, and S Shyam Sundar.
  In Companion of the 2025 ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp), Oct 2025.
  Paper

• Biosensing-driven Cognitive Augmentation: Real-time Mind Wandering Detection and Robot Intervention.
  Hyobin Kim, Jinkwon Lee, Seoyeong Kim, Jangwon Lee, and Hayoung Oh.
  In Companion of the 2025 ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp), Oct 2025.

• Real-time Traffic Accident Anticipation with Feature Reuse.
  Inpyo Song and Jangwon Lee.
  In IEEE International Conference on Image Processing (ICIP), Sep 2025. (Poster, 39.6% acceptance rate).
  Paper  |  Project Page  |  arXiv  |  Code 

• Pawprint: Whose Footprints Are These? Identifying Animal Individuals by their Footprints.
  Inpyo Song, Hyemin Hwang, and Jangwon Lee.
  In IEEE International Conference on Image Processing (ICIP), Sep 2025. (Poster, 39.6% acceptance rate).
  Paper  |  Project Page  |  arXiv  |  Code 

• SplitStream: Foreground–Background Separation based 3D Gaussian Splatting Streaming Technique
  Hannah Cho, Hyemin Hwang, Jangwon Lee
  Journal of Broadcast Engineering (JBE), VOL. 30, NO. 6

• Cross-Racial Facial Emotion Recognition Generalization Using Domain-Adversarial Neural Networks
  Minjun Ju, Joonhyung Kwon, Hyemin Hwang, Hannah Cho, Jangwon Lee
  The Korean Institute of Broadcast and Media Engineers (KIBME) Summer Conference

• Real-Time Multi-Object Tracking with DeepSORT on Mobile GPU Platforms
  Jongmin Lee, Yeeun Kim, Jangwon Lee
  The Korean Institute of Broadcast and Media Engineers (KIBME) Summer Conference

• SplitStream: A Foreground-Background Separated 3D Gaussian Splatting Streaming Method
  Hyemin Hwang, Hannah Cho, Minjun Ju, Joonhyung Kwon, Jangwon Lee
  The Korean Institute of Broadcast and Media Engineers (KIBME) Summer Conference

• DualSplat: Streaming 3DGS via Foreground-Background Decomposition
  Hyemin Hwang, Hannah Cho, Jangwon Lee
  Asia Pacific International Conference on Information Science and Technology (APIC-IST)

• Designing a Programming Course Model Using AI-Based Pair Programming Techniques
  Hyemin Hwang, Jangwon Lee
  The Journal of Korean Association of Computer Education, 28(2), 13-22

• Anomaly Detection for People with Visual Impairments using an Egocentric 360-degree Camera.
  Inpyo Song, Sanghyeon Lee, Minjun Joo, and Jangwon Lee.
  In IEEE Winter Conference on Applications of Computer Vision (WACV), Feb 2025. (Poster, 37.8% acceptance rate).
  Paper  |  Project Page  |  arXiv  |  Code  

2024

• Action-conditioned Contrastive Learning for 3d Human Pose and Shape Estimation in Videos.
  Inpyo Song, Moonwook Ryu, and Jangwon Lee.
  Computer Vision and Image Understanding (CVIU), 2024. (impact factor = 4.3).
  Paper  |  Code

• Sftrack: A Robust Scale and Motion Adaptive Algorithm for Tracking Small and Fast Moving Objects.
  Inpyo Song and Jangwon Lee.
  In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Oct 2024. (Oral, Selected Submission).
  Paper  |  Project Page  |  arXiv  |  Code

• Teacher Response Entropy Based Adaptive Knowledge Distillation
  Inpyo Song and Jangwon Lee
  Journal of Korean Institute of Intelligent Systems (KIIS)

• Rule-based Zero-shot Video Anomaly Detection Using Object Detection and Semantic Segmentation
  Inpyo Song and Jangwon Lee
  Journal of Broadcast Engineering (JBE), Vol. 29, No. 6

• Motion-aware Heatmap Regression for Human Pose Estimation in Videos.
  Inpyo Song, Jongmin Lee, Moonwook Ryu, and Jangwon Lee.
  In International Joint Conference on Artificial Intelligence, (IJCAI), Aug 2024. (Poster+Oral, 14.0% acceptance rate).
  Paper  |  Code

• Video Question Answering for People with Visual Impairments using an Egocentric 360-degree Camera.
  Inpyo Song, Minjun Joo, Joonhyung Kwon, and Jangwon Lee.
  In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshop on First Joint Egocentric Video (EgoVis), Jun 2024.
  arXiv

• Smooth and Consistent Pose Estimation with Intersections Module
  Jongmin Lee*, Inpyo Song*(Equal Contribution), Moonwook Ryu, Jangwon Lee
  The Korean Institute of Broadcast and Media Engineers (KIBME) Summer Conference

• SMPReID-37: A New Dataset for Smartphone Re-identification
  Jongmin Lee, Hannah Cho, Hyemin Hwang, Minjun Joo, Joonhyung Kwon
  The Korea Institute of Broadcast and Media Engineers (KIBME) Summer Conference

2023

• Scale and Motion Adaptive Multi-Object Tracking Algorithm for Unmanned Aerial Vehicles.
  Inpyo Song and Jangwon Lee.
  In International Conference on Robot and Human Interactive Communication (RO-MAN), Late Breaking Report, Aug 2023.

• Multi-task Learning in Facial Kinship Verification
  Inpyo Song and Jangwon Lee
  The Korean Institute of Broadcast and Media Engineers (KIBME) Summer Conference

2022

• Anomaly Detection for Visually Impaired People using a 360 degree Wearable Camera.
  Dong in Kim and Jangwon Lee.
  In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops on VizWiz Grand Challenge: Describing Images and Videos Taken by Blind People, Jun 2022. (Oral).

2021

• A Sample Weighting and Score Aggregation Method for Multi-Query Object Matching.
  Jangwon Lee, Gang Qian, and Allison Beach.
  In IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), Nov 2021. (Oral).

Publications (Up to 2020)

2019

• Observing Pianist Accuracy and Form with Computer Vision.
  Jangwon Lee, Bardia Doosti, Yupeng Gu, David Cartledge, David J. Crandall, and Christopher Raphael.
  In IEEE Winter Conference on Applications of Computer Vision (WACV), Jan 2019. (Poster+Oral, 39.0% acceptance rate).

2018

• Forecasting Hand and Object Locations in Future Frames.
  Chenyou Fan, Jangwon Lee, and Michael S Ryoo.
  In European Conference on Computer Vision (ECCV) Workshop on Anticipating Human Behavior, Sep 2018.

• Human-Drone Interaction: Drone Delivery & Services for Social Events.
  Haodan Tan, Jangwon Lee, and Gege Gao.
  In ACM SIGCHI Conference on Designing Interactive Systems (DIS), Works-in-Progress, Jun 2018.

2017

• Learning Robot Activities from First-Person Human Videos using Convolutional Future Regression.
  Jangwon Lee and Michael S. Ryoo.
  In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops on Deep Learning for Robotic Vision (DLRV), Jul 2017. (Spotlight Presentation). Best paper award!

• Identifying First-Person Camera Wearers in Third-Person Videos.
  Chenyou Fan, Jangwon Lee, Mingze Xu, Krishna Kumar Singh, Yong Jae Lee, David J. Crandall, and Michael S. Ryoo.
  In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Jul 2017. (Poster, 29.0% acceptance rate).

• Real-time, Cloud-based Object Detection for Unmanned Aerial Vehicles
  Jangwon Lee, Jingya Wang, David Crandall, Selma Šabanović, and Geoffrey Fox
  IEEE International Conference on Robotic Computing (IRC)

2009

• Environment Adaptive 3d Object Recognition and Pose Estimation by Cognitive Perception Engine.
  Hyunjun Kim, Jangwon Lee, and Sukhan Lee.
  In IEEE International Symposium on Computational Intelligence in Robotics and Automation (CIRA), Dec 2009.

2007

• Study on Behavioral Personality of a Service Robot to Make More Convenient to Customer.
  Jangwon Lee, Dongwook Shin, HunSue Lee, and Sukhan Lee.
  In IEEE International Symposium on Robot and Human interactive Communication (RO-MAN), Aug 2007.

• Information Integration and Mission Selection to Accomplish Dependable Perception for Service Robot.
  Seung-Min Baek, Jangwon Lee, Hunsue Lee, Dongwook Shin, and Sukhan Lee.
  In IEEE International Conference on Advanced Robotics (ICAR), Aug 2007.

• Security Service Robot in Ubiquitous Environment based on Cognitive Robotic Engine.
  Hunsue Lee, Jangwon Lee, Jaewoong Kim, and Sukhan Lee.
  In International Conference of Ubiquitous Information Technology and Applications (ICUT), Feb 2007.

2006

• Robot Personality from Perceptual Behavior Engine: An Experimental Study.
  Dongwook Shin, Jangwon Lee, Hun-Sue Lee, Sukhan Lee, Young-Jo Cho, and Su-Young Chi.
  In International Conference on Ubiquitous Robots and AmbientIntelligence (URAI), Oct 2006.