The 3D world around us is composed of a rich variety of objects: buildings, bridges, trees, cars, rivers, and so forth, each with distinct appearance, morphology, and function. Giving machines the ability to precisely segment and label these diverse objects is of key importance to allow them to interact competently within our physical world, for applications such as scene-level robot navigation, autonomous driving, and even large-scale urban 3D modeling, which is critical for the future of smart city planning and management.

Over the past years, remarkable advances in techniques for 3D point cloud understanding have greatly boosted performance. Although these approaches achieve impressive results for object recognition and semantic segmentation, almost all of them are limited to extremely small 3D point clouds, and are difficult to be directly extended to large-scale point clouds. [Bilibili Live] [YouTube Live]

The 2nd Challenge on Large Scale Point-cloud Analysis for Urban Scenes Understanding (Urban3D) at ECCV 2022 aims to establish new benchmarks for 3D semantic and instance segmentation on urban-scale point clouds. In particular, we prime the challenge with both SensatUrban and STPLS3D datasets. SensatUrban consists of large-scale subsections of multiple urban areas in the UK. With the high quality of per-point annotations and the diverse distribution of semantic categories. STPLS3D is composed of both real-world and synthetic environments which cover more than 17 km2 of the city landscape in the U.S. with up to 18 fine-grained semantic classes and 14 instance classes. These two datasets are complementary to each other and allow us to explore a number of key research problems and directions for 3D semantic and instance learning in this workshop. We aspire to highlight the challenges faced in 3D segmentation on extremely large and dense point clouds of urban environments, sparking innovation in applications such as smart cities, digital twins, autonomous vehicles, automated asset management of large national infrastructures, and intelligent construction sites. We hope that our datasets, and this workshop could inspire the community to explore the next level of 3D learning. Specifically, We encourage researchers from a wide range of background to participate in our challenge, the topics including but not limited to:

  • Semantic segmentation of large-scale 3D point clouds.
  • Instance segmentation of 3D point clouds.
  • Weakly supervised learning in 3D point clouds analysis.
  • Learning from imbalanced 3D point clouds.
  • 3D point cloud acquisition & visualization.
  • 3D object detection & reconstruction.
  • Semi-/weak-/un-/self- supervised learning methods for 3D point clouds.
We will be hosting 2 invited speakers and holding 2 parallel challenges (i.e., semantic and instance segmentation), and 1 panel discussion session for the topic of point cloud segmentation. More information will be provided as soon as possible.

Call for Contributions

Urban3D Challenges@ECCV'2022

The Urban3D Challenges are hosted on Codalab, and can be found at:

We are thankful to our sponsor for providing the following prizes. The prize award will be granted to the Top 3 individuals and teams for Each Challenge Track on the leaderboard that provide a valid submission.

    • 1st Place:
$1,500 USD courtesy of
    • 2nd Place:
$1,000 USD courtesy of
    • 3rd Place:
$500 USD courtesy of

Important Dates

Workshop Proposal Accepted March 26, 2022
Competition Starts May 12, 2022
Competition Ends October 10, 2022 (23:59 Pacific time)
Notification to Participants October 15, 2022
Finalized Workshop Program (Half Day) October 23, 2022 (9:00-13:00 IDT (UTC+3))

Preliminary Program Outline

09:00-09:05 Welcome Introduction
09:05-09:50 Invited Talk (Talk 1)
09:50-10:35 Invited Talk (Talk 2)
10:35-11:20 Invited Talk (Talk 3)
11:20-11:25 Coffee break
11:25-11:40 Winner Talk 1 (Track 1)
11:40-11:55 Winner Talk 2 (Track 1)
11:55-12:10 Winner Talk 3 (Track 1)
12:10-12:25 Winner Talk 1 (Track 2)
12:25-12:40 Winner Talk 2 (Track 2)
12:40-12:55 Winner Talk 3 (Track 2)
12:55-13:00 Closing Remarks

Invited Keynote Speakers

Randall W. Hill, Jr.
USC Institute for Creative Technologies

The Quest to Build the Holodeck

Biography (click to expand/collapse)

Randall W. Hill, Jr. became the executive director of the USC Institute for Creative Technologies in 2006. A leader in understanding how classic storytelling and high-tech tools can create meaningful learning experiences, Hill steers the institute’s exploration of how virtual humans, mixed reality worlds, advanced computer graphics, dramatic films, social simulations and educational videogames can augment more traditional methods for imparting lessons. He oversees a diverse team of scientists, storytellers, artists and educators as they pioneer and evaluate new ways to deliver effective teaching and training in areas including leadership, cultural awareness, negotiation and mental health treatment and assessment.

He is a research professor in the computer science department at the USC Viterbi School of Engineering. His research focus is on using intelligent tutoring systems and virtual humans to augment immersive learning environments. Hill’s career at USC began in 1995 at the USC Information Sciences Institute where he worked on the development of models of human behavior and decision-making for real-time simulation environments. He joined the USC Institute for Creative Technologies in 2000 as a senior scientist. Prior to joining USC, Hill served as a group supervisor and the work area manager for network automation in the Deep Space Network Advanced Technology Program at NASA’s Jet Propulsion Laboratory.

Hill graduated with a Bachelor of Science degree from the United States Military Academy at West Point and earned his M.S. and Ph.D. degrees in computer science from the University of Southern California.

SHI Wen-zhong
The Hong Kong Polytechnic University

Building 3D Urban Models by Mobile Mapping System

Biography (click to expand/collapse)

Professor John Shi is currently the Director of Otto Poon Charitable Foundation Smart Cities Research Institute of PolyU, Director of PolyU-Shenzhen Technology and Innovation Research Institute (Futian), Chair Professor in Geographic Science and Remote Sensing, and Director of Joint Research Laboratory on Spatial Information of PolyU and Wuhan University. He is Academician of International Eurasian Academy of Sciences and Fellow of Academy of Social Sciences (UK). He earned his doctoral degree from University of Osnabruck in Vechta, Germany in 1994. A Fellow of Royal Institution of Chartered Surveyors and Hong Kong Institute of Surveyors, Professor Shi also serves as President of International Society for Urban Informatics and Editor-in-Chief of International Journal Urban Informatics.

His research covers urban informatics for smart cities, geo-informatic science and remote sensing, artificial-intelligence-based object extraction and change detection from satellite imagery, intelligent analytics and quality control for spatial big data, and mobile mapping and 3-D modelling based on LiDAR and remote sensing imagery. He has published almost 300 research articles in journals indexed by Web of Science and 20 books. He is among the worldly top 2% cited researchers according to the standardized citation indicators published by Elsevier BV and scholar in Stanford University. He has 34 patents grants and 33 other patent applications filed.

Professor Shi has won Natural Science Award, China’s highest award for fundamental research, in 2007; Distinguished Scholar Prize by CPGIS, Gold Medal in Geneva Invention Expo, and Smart 50 Awards (US) in 2021; Founder’s Award by International Spatial Accuracy Research Association in 2020; China’s Science and Technology Progress Award in Surveying and Mapping (Grand Award) in 2017; Wang Zhizhuo Award by International Society of Photogrammetry and Remote Sensing in 2012; and ESRI Award for Best Scientific Paper in Geo-informatic Science by American Society of Photogrammetry and Remote Sensing in 2006.

Virtual Humans — From appearance to behaviour

Biography (click to expand/collapse)

Gerard Pons-Moll is a Professor at the University of Tübingen, at the department of Computer Science. He is also the head of the Emmy Noether independent research group "Real Virtual Humans", senior researcher at the Max Planck for Informatics (MPII) in Saarbrücken, Germany, and faculty at the IMPRS-IS (International Max Planck Research School - Intelligent Systems in Tübingen) and faculty at Saarland Informatics Campus. His research lies at the intersection of computer vision, computer graphics and machine learning -- with special focus on analyzing people in videos, and creating virtual human models by "looking" at real ones. His research has produced some of the most advanced statistical human body models of pose, shape, soft-tissue and clothing (which are currently used for a number of applications in industry and research), as well as algorithms to track and reconstruct 3D people models from images, video, depth, and IMUs.

His work received several awards such as the Emmy Noether grant’18, German Pattern Recognition Award’19, and best papers and nominations at CVPR’21, CVPR’20, IVA’21, 3DV’22, 3DV’18, Eurographics’17 and BMVC’2013.


Winner Award sponsored by

Track 1: 3D Semantic Segmentation of Urban-scale Point Clouds.

  • 1st Place:   USTC SpaceAI
    Smooth Denoising for 3D Semantic Segmentation
    Xinjun Li, Jiahao Lu, Yihan Chen, Jiacheng Deng, Chuxin Wang

  • 2nd Place:   B•B•G
    Mixed Point-based and Voxel-based Networks for 3D Point Cloud Processing
    Zeqiang Wei, Kai Jin, Angulia Yang, Mingzhi Gao, Kuan Song, Xiuzhuang Zhou

  • 3rd Place:   Deep Bit Lab
    Enhanced Point Transformer
    Xu Yan, Jiantao Gao, Zhuo Li, Zhen Li, Yan Peng, Shuguang Cui

Track 2: 3D Instance Segmentation of Urban-scale Point Clouds.

  • 1st Place:   B•B•G
    Mixed Point-based and Voxel-based Networks for 3D Point Cloud Processing
    Zeqiang Wei, Kai Jin, Angulia Yang, Mingzhi Gao, Kuan Song, Xiuzhuang Zhou

  • 2nd Place:   Mask3D People
    Mask3D for 3D Semantic Instance Segmentation
    Jonas Schult, Francis Engelmann, Alexander Hermans, Or Litany, Siyu Tang and Bastian Leibe

  • 3rd Place:   Lidiar-3D
    A dual-function point cloud segmentation network
    Tengping Jiang, Qinyu Zhang, Lin Zhao, Yuan Zhao, Zequn Zhang


Qingyong Hu
University of Oxford
Meida Chen
University of Southern California - Institute for Creative Technologies
Ta-Ying Cheng
University of Oxford
Sheikh Khalid
Sensat LTD.
Bo Yang
The Hong Kong Polytechnic University

Ronarld Clark
Imperial College London
Jiahui Chen
Sun Yat-sen University
Leying Zhang
Sun Yat-sen University
Rongkun Yang
Sun Yat-sen University
Yulan Guo
National University of Defense Technology

Aleš Leonardis
University of Birmingham
Niki Trigoni
University of Oxford
Andrew Markham
University of Oxford

Workshop sponsored by:

Previous years' workshops: