Sungmin Woo

I am a Ph.D candidate at Yonsei University, Seoul, Korea.

My research mainly focuses on 3D computer vision tasks such as depth estimation, point cloud processing, neural rendering, and motion prediction.

Please feel free to contact me if you have any questions or suggestions :)

Email / CV / Google Scholar

Yonsei University B.S in EE Mar. 16 - Feb. 20	Yonsei University Ph.D in EE Mar. 20 - Feb. 25

Selected Publication

	ProDepth: Boosting Self-Supervised Multi-Frame Monocular Depth with Probabilistic Fusion Sungmin Woo, Wonjoon Lee, Woojin Kim, Dogyoon Lee, Sangyoun Lee European Conference on Computer Vision (ECCV) , 2024 Paper / arXiv / Code / Project Page We propose a novel framework called ProDepth, which effectively addresses the mismatch problem caused by dynamic objects using a probabilistic approach.
	FIMP: Future Interaction Modeling for Multi-Agent Motion Prediction Sungmin Woo, Minjung Kim, Donghyeong Kim, Sangyoun Lee IEEE International Conference on Robotics and Automation (ICRA), 2024 Paper / arXiv / Code We propose Future Interaction modeling for Motion Prediction (FIMP), which captures potential future interactions in an endto-end manner.
	Multi-scale Structural Graph Convolutional Network for Skeleton-based Action Recognition Sungjun Jang, Heansung Lee, Woojin Kim, Jungho Lee, Sungmin Woo, Sangyoun Lee IEEE Transactions on Circuits and Systems for Video Technology (T-CSVT), 2024 Paper / arXiv / Code We propose the multi-scale structural graph convolutional network (MSS-GCN) for skeleton-based action recognition.
	MKConv: Multidimensional Feature Representation for Point Cloud Analysis Sungmin Woo, Dogyoon Lee, Sangwon Hwang, Woojin Kim, Sangyoun Lee Pattern Recognition (PR), 2023 Paper / arXiv / Code We propose a novel convolution operation for point cloud processing, introducing a multidimensional feature representation.
	Leveraging Spatio-Temporal Dependency for Skeleton-Based Action Recognition Jungho Lee, Minhyeok Lee, Suhwan Cho, Sungmin Woo, Sungjun Jang, Sangyoun Lee IEEE/CVF International Conference on Computer Vision (ICCV), 2023 Paper / arXiv / Code We propose a novel Spatio-Temporal Curve Network (STC-Net) for skeleton-based action recognition, which consists of spatial modules with an spatio-temporal curve (STC) module and graph convolution with dilated kernels.
	MSV-RGNN: Multiscale Voxel Graph Neural Network for 3D Object Detection Wonjoon Lee, Sungmin Woo, Donghyeong Kim, Sangyoun Lee IEEE International Conference on Image Processing (ICIP), 2023 Paper / arXiv / Code We propose MSV-RGNN, a two-stage 3D object detector that utilizes multiple sets of graphs across all scales via a multiscale-voxel-graph RoI pooling module.
	Detection-Identification Balancing Margin Loss for One-Stage Multi-Object Tracking Heansung Lee, Suhwan Cho, Sungjun Jang, Jungho Lee, Sungmin Woo, Sangyoun Lee IEEE International Conference on Image Processing (ICIP), 2022 Paper / arXiv / Code We present a Detection-Identification balancing Margin loss on one-stage MOT for suppressing negative transfer effect to achieve balanced performance of detection and re-ID.
	Regularization Strategy for Point Cloud via Rigidly Mixed Sample Dogyoon Lee, Jaeha Lee, Junhyeop Lee, Hyeongmin Lee, Minhyeok Lee, Sungmin Woo, Sangyoun Lee IEEE/CVF Computer Vision and Pattern Recognition (CVPR), 2021 Paper / arXiv / Code We propose a novel data augmentation method called Rigid Subset Mix (RSMix) which generates virtual mixed samples by replacing part of the sample with shape-preserved subsets from another sample.
	Lidar depth completion using color-embedded information via knowledge distillation Sangwon Hwang, Junhyeop Lee, Woojin Kim, Sungmin Woo, Kyungjae Lee, Sangyoun Lee IEEE Transactions on Intelligent Transportation Systems (T-ITS), 2021 Paper / arXiv / Code We present a depth completion framework consisting of depth and edge CNNs with transferring of knowledge.
	AIBM: Accurate and instant background modeling for moving object detection Woojin Kim, Sangwon Hwang, Junhyeop Lee, Sungmin Woo, Sangyoun Lee IEEE Transactions on Intelligent Transportation Systems (T-ITS), 2021 Paper / arXiv / Code We propose a novel background modeling method for moving object detection based on inpainting and enhancing the resolution using a coarse-to-fine strategy.
	Ghost graph convolutional network for skeleton-based action recognition Sungjun Jang, Heansung Lee, Suhwan Cho, Sungmin Woo, Sangyoun Lee IEEE International Conference on Consumer Electronics-Asia (ICCE-Asia)), 2021 Paper / arXiv / Code We propose a ghost graph convolution for skeleton-based action recognition.
	False Positive Removal For 3D Vehicle Detection with Penetrated Point Classifier Sungmin Woo, Sangwon Hwang, Woojin Kim, Junhyeop Lee, Dogyoon Lee, Sangyoun Lee IEEE International Conference on Image Processing (ICIP), 2020 Paper / arXiv / Code We propose a novel post-processing method to remove false positives in 3D vehicle detection utilizing the characteristics of the LiDAR sensor itself.

Preprint

This website's source code is borrowed from jonbarron's website.

Last updated July 2024.