1st Workshop on Scene Understanding
in Unstructured Environments

In conjunction with the DAGM GCPR 2021, September 28


Registration is now open at virtual21.dagm-gcpr.de

Important Dates

    TAS500 Dataset available for Download: April 15, 2021
     Outdoor Semantic Segmentation Challenge Deadline: August 15, 2021
     Workshop Date: September 28, 2021 (Registration)

     Visit the Codalab Competition Page for more information.


Competition ResultsNew!

The Outdoor Semantic Segmentation Challenge concluded on August 15 and the 3 best-performing are invited to present their results at the 1st Workshop on Scene Understanding in Unstructured Environments on September 28, 2021. The following three teams are invited for participation:

    1st place: User ACVLabNCKU with a mIoU score of 67.50% from the Advanced Computer Vision Lab at the National Cheng Kung University in Taiwan.
     2nd place: User DPSoler with a mIoU score of 66.32% from the Institute of Mathematics and Statistics, University of São Paulo in Brazil. A video of the presentation and a technical report for their solution is also available.
     3rd place: User andre.vmatias with a mIoU score of 65.87% from the Federal University of Santa Catarina in Brazil. A video of the presentation and a technical report for their solution is also available.

We want to thank all active participants for taking part in this competition. The leaderboard containing all submissions is available here. We invite all participants to visit the virtual workshop on September 28.

The registration for the virtual DAGM GCPR 2021 is now available at https://virtual21.dagm-gcpr.de. Don't forget to select the "Workshop on Scene Understanding in Unstructured Environments" as your "Interest" during registration. This allows us to estimate the event attendance and prepare accordingly. The registration for the virtual DAGM GCPR 2021 is free.



Technical Report SubmissionNew!

Finalists of the Outdoor Semantic Segmentation Challenge can submit a written technical report. The submissions will be processed via CMT, follow this link to the CMT submission page. We ask submissions follow the same format guidelines as used by the regular submissions for the DAGM GCPR 2021. Please check the "Submission" section of the GCPR website for more details. Our review process differs by being single blind. This means, that we do not require to anonymize the author and his affiliation in the submission.

Here is a brief overview about the general form of the written technical report:

    • The technical report has to be written in English and has to conform with the Springer LNCS format, using the Springer's proceedings LaTeX template. It is also available in Overleaf, but dagmgcpr2021.tex from the author kit must be used.
     • The technical report length is limited to 12 pages excluding references. It is not allowed to modify the margins, font size, or page layout of the template.
     • All technical reports have to be submitted as a single PDF file. The authors are requested to make sure that all fonts are embedded in the PDF file.


Outdoor Semantic Segmentation Challenge

As part of the first workshop on scene understanding in unstructured environments we introduce the outdoor semantic segmentation challenge. The task is to develop an algorithm for a fine-grained vegetation and surface detection in an unstructured outdoor environment. The training data is provided by the TAS500 dataset. Participants are invited to send their predictions on the test set. All submissions are added onto our public leaderboard.

The best submissions will be invited so present their solution as part of the workshop at the DAGM GCPR 2021.



Evaluation

The submitted predictions are going to be evaluated on the TAS500 test split. For evaluation we will use the mean Intserection over Union (mIoU) metric. In addition, we will also display the novel Boundarry Jaccard (BJ) metric for each submission. The mean Boundary Jaccard (mBJ) focuses on evaluating pixels along class boundaries. This can be more relevant for fine-grained semantic segmentation tasks such as the scenes shown in the TAS500 dataset. Both the mIoU and the mBJ are highly correlated metrics. The final ranking on the leaderboard is done using the mIoU metric.



Codalab Competition

The Outdoor Semantic Segmentation Challenge is hosted on Codalab. CodaLab is an open-source web-based platform to collaborate on challenging resarch datasets. The public leaderboard and the automatic submission evaluation for this workshop challenge is hosted there.

We use Codalab's competition forum for discussions regarding the competition.



Download

The challenge uses version 1.1 of the TAS500 dataset. The TAS500v1.1 dataset is available for download here.




More information on the TAS500 dataset is available on the project page.



Submission

All submissions are handled in the Codalab Competition. For more information visit the Participate section on Codalab.