3rd ScanNet Indoor Scene Understanding Challenge

CVPR 2021 Workshop

June 20, 2021



Introduction

3D scene understanding for indoor environments is becoming an increasingly important area. Application domains such as augmented and virtual reality, computational photography, interior design, and autonomous mobile robots all require a deep understanding of 3D interior spaces, the semantics of objects that are present, and their relative configurations in 3D space.

We present the first comprehensive challenge for 3D scene understanding of entire rooms at the object instance-level with 5 tasks based on the ScanNet dataset. The ScanNet dataset is a large-scale semantically annotated dataset of 3D mesh reconstructions of interior spaces (approx. 1500 rooms and 2.5 million RGB-D frames). It is used by more than 480 research groups to develop and benchmark state-of-the-art approaches in semantic scene understanding. A key goal of this challenge is to compare state-of-the-art approaches operating on image data (including RGB-D) with approaches operating directly on 3D data (point cloud, or surface mesh representations). Additionally, we pose both object category label prediction (commonly referred to as semantic segmentation), and instance-level object recognition (object instance prediction and category label prediction). We propose five tasks that cover this space:

  • 2D semantic label prediction: prediction of object category labels from 2D image representation
  • 2D semantic instance prediction: prediction of object instance and category labels from 2D image representation
  • 3D semantic label prediction: prediction of object category labels from 3D representation
  • 3D semantic instance prediction: prediction of object instance and category labels from 3D representation
  • Scene type classification: classification of entire 3D room into a scene type

New This Year - Data Efficient Challenge!

In the data efficient challenge, training is conducted on Limited Scene Reconstructions (LR) or Limited Scene Annotations (LA), for the tasks of 3D Semantic Segmentation, Instance Segmentation and Object Detection.

For each task, challenge participants are provided with prepared training, validation, and test datasets, and automated evaluation scripts. In addition to the public train-val-test split, benchmarking is done on a hidden test set whose raw data can be downloaded without annotations; in order to participate in the benchmark, the predictions on the hidden test set are uploaded to the evaluation server, where they are evaluated. Submission is restricted to submissions every two weeks to avoid finetuning on the test dataset. See more details at http://kaldir.vc.in.tum.de/scannet_benchmark/documentation if you would like to participate in the challenge. The evaluation server leaderboard is live at http://kaldir.vc.in.tum.de/scannet_benchmark/. See the new data efficient documentation and leaderboard!.


2D semantic label prediction

2D semantic instance prediction

3D semantic label prediction

3D semantic instance prediction

TBDTBD


Schedule

PDTCEST
Welcome and Introduction 8:50am - 9:00am5:50pm - 6:00pm
Invited Speaker 1 9:00am - 9:30am6:00pm - 6:30pm
Invited Speaker 2 9:30am - 10:00am6:30pm - 7:00pm
Winner Talk 1 10:00am - 10:10am7:00pm - 7:10pm
Winner Talk 2 10:10am - 10:20am7:10pm - 7:20pm
Winner Talk 3 10:20am - 10:30am7:20pm - 7:30pm
Invited Speaker 3 10:30am - 11:00am7:30pm - 8:00pm
Invited Speaker 4 11:00am - 11:30am8:00pm - 8:30pm
Invited Speaker 5 11:30am - 12:00pm8:30pm - 9:00pm
Panel Discussion and Conclusion 12:00pm - 12:30pm9:00pm - 9:30pm


Invited Speakers


Niloy Mitra is a Professor of Geometry Processing in the Department of Computer Science, University College London (UCL). He received his MS and PhD in Electrical Engineering from Stanford University under the guidance of Leonidas Guibas and Marc Levoy, and was a postdoctoral scholar with Helmut Pottmann at Technical University Vienna. His research interests include shape analysis, computational design and fabrication, and geometry processing. Niloy received the 2013 ACM Siggraph Significant New Researcher Award for "his outstanding work in discovery and use of structure and function in 3D objects" (UCL press release), the BCS Roger Needham award (BCS press release) in 2015, and the Eurographics Outstanding Technical Contributions Award in 2019. He received the ERC Starting Grant on SmartGeometry in 2013. His work has twice been featured as research highlights in the Communications of the ACM, twice been selected by ACM Siggraph/Siggraph Asia (both in 2017) for press release as research highlight.


Bastian Leibe is a full Professor of Computer Science at RWTH Aachen University, Germany, where he leads the Computer Vision group. He holds an MSc degree from the Georgia Institute of Technology (1999), a Diploma degree from the University of Stuttgart (2001), and a PhD from ETH Zurich (2004), all three in Computer Science. His main research interests are in computer vision and machine learning for dynamic visual scene understanding, encompassing object recognition, tracking, segmentation, and 3D reconstruction. He has published over 150 articles in peer-reviewed journals and conferences. Over the years, he has received several awards for his research work, including the CVPR Best Paper Award in 2007, the DAGM Olympus Prize in 2008, and the U.V. Helava Award in 2012. In 2012, he was awarded a European Research Council (ERC) Starting Grant, and in 2017 an ERC Consolidator Grant. He has been Program Chair for ECCV 2016 and Area Chair and program committee member for all major computer vision conferences.


Charles Qi is a research scientist at Waymo LLC. Previously, he was a postdoctoral researcher at Facebook AI Research (FAIR) and received his Ph.D. from Stanford University (Stanford AI Lab and Geometric Computation Group), advised by Professor Leonidas J. Guibas. Prior to Stanford, he received his B.Eng. from Tsinghua University. His research focuses on deep learning, computer vision and 3D. He has developed novel deep learning architectures for 3D data (point clouds, volumetric grids and multi-view images) that have wide applications in 3D object classification, object part segmentation, semantic scene parsing, scene flow estimation and 3D reconstruction; such deep architectures have been well adopted by both academic and industrial groups across the world.


Lu Fang is an Associate Professor at Tsinghua University. She received her Ph.D from the Hong Kong University of Science and Technology in 2011, and her B.E. from Univ. of Science and Technology of China in 2007, respectively. Prof. Fang's research interests include computational imaging and 3D vision. Prof. Fang's research has been recognized by the NSFC Excellent Young Scholar Award, Multimedia Rising Star Award in ICME 2019, Best Student Paper Award in ICME 2017, and more. Prof. Fang is an currently IEEE Senior Member, and serving as an Associate Editor for IEEE TMM and TCSVT.


Saining Xie is a research scientist at Facebook AI Research (FAIR). He received a PhD in computer science from the University of California San Diego in 2018, advised by Zhuowen Tu. Prior to that, he received his B.S. from Shanghai Jiao Tong University in 2013. He has broad research interests in computer vision and deep learning, with a focus on representation learning for visual recognition in both 2D and 3D. He is a recipient of the Google Ph.D. fellowship in 2017 and the Marr Prize Honorable Mention award at ICCV 2015.


Organizers

Angela Dai
Technical University of Munich
Angel X. Chang
Simon Fraser University
Manolis Savva
Simon Fraser University
Matthias Niessner
Technical University of Munich

Acknowledgments

Thanks to visualdialog.org for the webpage format.