Spring 2022

In this semester, we mainly worked on our novel method MOST: MOdal Segmentation and Tracking which marries object instance recognition and semantic segmentation in a single network, achieving state-of-the-art performance for both 3D and 4D lidar panoptic segmentation. Our work is also submitted to a top computer vision conference.

  • First Presentation Slides [Link]
  • First Presentation Video [Link]
  • Second Presentation Slides [Link]
  • Second Presentation Video [Link]
  • Final Presentation Slides [Link]
  • Final Presentation Video [Link]