Embedded Files
지능형 미디어 연구실
Seoul National University of Science and Technology
Welcome!
Welcome!
Welcome to visit IMRL website!
IMRL is at SeoulTech Mirae Hall Rm. 305.
Our group is solving diverse world problems including research on computer vision and time-series data analysis. We are working towards the development of efficient, robust, and reliable systems with deep learning and artificial intelligence.
Note: IMRL is actively looking for self-motivated graduate and undergraduate students to join our lab!
성실히 인공지능 관련한 공부와 연구를 할 준비가 되어있는 학생 연구원을 모집합니다.
NEWS:
Congratulations!🎉 2025 한국멀티미디어학회 춘계학술발표대회 우수발표상 수상, 학부연구생 정재찬, 김준수
"멀티모달 기반의 지식 증류를 활용한 시계열 데이터의 분류 성능 향상 연구"
"DCR KD: Dynamic Class Relation Knowledge Distillation for Semantic Segmentation with the Frontal-viewing Camera of Limited Field of View in an Internet of Things Environment" has been accepted at IEEE IoTJ! [pdf]
(Google Scholar - Computer Systems Top 1 publication)
"Ground Reaction Force Estimation via Time-aware Knowledge Distillation" has been accepted at IEEE IoTJ! [pdf] [arxiv] [linkedin]
(Google Scholar - Computer Systems Top 1 publication)
3D Gen AI Paper "MT3D" has been accepted at CVPR 2025 workshop! (AI for Content Creation Workshop)
"Intra-class Patch Swap for Self-Distillation" has been accepted at Neurocomputing! [pdf] [arxiv]
(Google Scholar - Artificial Intelligence Top 8 publication)
Spotlight on our collaborator, GML at ASU (News) [linkedin]
"Role of Mixup in Topological KD" has been published at IEEE Sensors Journal! [pdf] [linkedin]
"Robustness of Topological Persistence for Wearable Sensor Data" has been published at EPJ Data Science! [pdf]
3D Gen AI Paper "MT3D" for WACV 2025 has been published! [project]
(Google Scholar - computer vision & pattern recognition Top 9 publication) ASU AME
2 NeurIPS workshop papers have been published!
Recent Research Topics!
Wearable Sensor Data Analysis
Wearable Sensor Data Analysis
For healthcare (disease diagnosis/prediction, injury prevention), we analyze wearable sensors (IMUs) as well as Insole pressure data, and develop efficient and robust deep learning models.
Cognitive Neural Framework
Cognitive Neural Framework
Learning Debiased and Interpretable Representations via Explanable AI (XAI).
Semantic Segmentation
Semantic Segmentation
Semantic Segmentation for autonomous driving and overcoming limited field-of-view issues.
Improving accuracy with a lightweight model.
Multimodal Data Analysis with AI
Multimodal Data Analysis with AI
By leveraging topological feature and multimodalities, an improved and efficient model is developed.
Generative AI
Generative AI
Improving performance of models by leveraging gen AI. (T2I, T23D, etc.)
Improving fidelity with better/controllable gen AI.
Reliable AI
Reliable AI
Improving performance of reliability as well as accuracy for AI models.
On-going project is related to multimodality, multidisciplinary AI, and generative AI models.
Have any question?
Have any question?
If you have any questions and professional contact, please send me an email ([ejeon6@seoultech.ac.kr]) for further details and discussion.
Page updated
Google Sites
Report abuse