JunYoung Gwak
Papers
6
Total Citations
384
H-Index
5
About
JunYoung Gwak is a researcher at the forefront of 3D scene understanding, spatial perception, and deep learning for robotics and augmented reality applications. His work focuses on developing novel neural network architectures and datasets that enable machines to perceive and interact with three-dimensional environments more effectively. Gwak's most notable contribution is his pioneering work on Minkowski Convolutional Neural Networks, which introduced 4D spatio-temporal convolutions for processing 3D video data — a breakthrough that has garnered over 124 citations and fundamentally advanced how robotics systems handle sequential depth and LiDAR inputs. He has also made significant strides in 3D object detection through his Generative Sparse Detection Networks (115 citations), addressing the unique challenges posed by sparse point cloud data. A recurring theme in Gwak's research is enabling robots to operate safely among humans. His JRDB dataset (117 citations) provides a rich multimodal benchmark for egocentric human perception, while his JRMOT framework advances real-time 3D multi-object tracking for autonomous navigation. His earlier work on DeformNet further demonstrated his breadth, tackling 3D shape reconstruction from single images. Collectively, Gwak's contributions position him as a key figure in bridging 3D deep learning with real-world robotic applications.
Research Focus
Key Achievements
Top Papers
- 14D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks124 citations · 2019
- 2
- 3Generative Sparse Detection Networks for 3D Single-Shot Object Detection115 citations · 2020
- 4JRMOT: A Real-Time 3D Multi-Object Tracker and a New Large-Scale Dataset17 citations · 2020
- 5Generative Sparse Detection Networks for 3D Single-shot Object Detection6 citations · 2020
- 6