Computer Vision in the Built Environment Call for Papers

Computer Vision in the Built Environment
for the Design, Construction, and Operation of Buildings

Dear Colleagues,

We invite you to participate in the first Scan-to-BIM challenge
focused on key problems when converting 3D point cloud data obtained
using lidar, photogrammetry, or depth map cameras to Building
Information Models (BIMs). For maintenance, retrofitting, or
renovation AEC entities could greatly benefit from access to BIMs of
their facilities. Furthermore, asset owners could use BIMs for
operations and maintenance as BIMs enable storing and having access to
attributes such as maintenance history, material type, and
manufacturer specifications. The challenge will include two tasks: (I)
Floorplan Reconstruction and (II) 3D Building Model
Reconstruction. For details on the two tasks please see below.

Top participants will be invited to present in the first Workshop and
Challenge on Computer Vision in the Built Environment for the Design,
Construction and Operation of Buildings, which will also feature
invited talks and a panel discussion.

More information and dataset download can be found on the workshop
website, if you have seen an earlier version of this invitation,
please note that the validation and testing data have been made
available now:

https://cv4aec.github.io/

We look forward to seeing your participation in the challenge!

The Workshop Organizers

Iro Armeni, Postdoctoral Researcher, ETHZ

Erzhuo Che, Professor, CEE, Oregon State University

Yong Cho, Professor, CEE, Georgia Tech University 

Martin Fischer, Professor, CEE, Stanford University

Daniel Hall, Professor, CEE, ETHZ 

Jaehoon Jung, Professor, CEE, Oregon State University

Fuxin Li, Professor, CS, Oregon State University

Michael Olsen, Professor, CEE, Oregon State University

Marc Pollefeys, Professor, CS, ETHZ 

Silvio Savarese, Professor, CS, Stanford University

Yelda Turkan, Professor, CEE, Oregon State University 


** CHALLENGE INFORMATION **

Important Dates

    Challenge submission deadline: June 1st, 2021

    Notification to authors: June 10th, 2021

    Workshop day: June 20, 2021, Saturday. Day 2 of CVPR 2021




I. Floorplan reconstruction task

Dataset: For the floorplan reconstruction task, the training set
consists of the point clouds of 20 buildings with multiple floors
each, with a total of 49 point clouds, as well as corresponding
aligned 2D building models for each floor with multiple extracted
layers. These layers include: walls, curved walls, doors, windows,
stairs, and columns. The buildings have complicated floor structures
which could include dozens of rooms, curved walls and many doors on
each floor. The validation set contains 21 point clouds and their
corresponding 2D building models, and is now available. The testing
set contains another 21 point clouds. For the testing set, only the
point clouds will be made available, while the 2D building models will
not be made available and we will host a server to evaluate the
submissions.

Evaluation Metrics: The evaluation consists of both geometric and
topological metrics. For geometric metrics, the IoU of the floor area
will be measured, as well as the endpoint accuracy, and orientation of
walls, doors, and windows. We will also evaluate topological metrics
that measure whether the connectivity of the rooms match the ground
truth.

II. 3D building model reconstruction task

Dataset: For the 3D building model reconstruction task, the training
set consists of point clouds of 4 buildings with multiple floors, as
well as the corresponding 3D building model in AutoCAD DXF format. We
aim to release 2 more buildings for the validation set and 2 more as
the testing set. The evaluation on the testing set will also be hosted
on the evaluation server in a similar manner to the floorplan
reconstruction task.

Evaluation Metrics: The evaluation will be done manually by an expert panel.