1st Workshop on Scene Understanding
in Unstructured Environments
In conjunction with GCPR 2021, September 28 - October 1, 2021
In the last decade we have seen much research activity in scene understanding by the emergence of deep learning in computer vision. The vast majority of developed solutions are applied in urban and structured environments. Unfortunately, these methods often fail when used in unstructured environments. Unstructured environments provide their own challenges such as varying lighting conditions and weather conditions as well as more natural structures in the scene. These conditions require different model assumptions and robust algorithms to accurately extract relevant information.
We are hosting the first workshop on scene understanding in unstructured environments at the DAGM GCPR 2021 to combine the research efforts and offer a platform for an exchange of knowledge within the research community. We will discuss different approaches that will find their way into the next generation of vision solutions for unstructured environments.
The workshop offers a broad view of different research areas in scene understanding in unstructured environments. Topics of interest include both 2D and 3D scene understanding in unstructured environments. We are also interested in advanced data augmentation and domain adaptation techniques to enhance the otherwise low availability of data for most unstructured environments. The use of multiple sensors in multimodal scene understanding solutions and their application in different domains are also of interest.
As part of this workshop we propose the oudoor semantic segmentation challenge on the recently released TAS500 dataset. Participants are invited to submit their proposed method to our public leaderboard and to present their method at the DAGM GCPR 2021 conference workshop in Bonn, Germany. You can find more out about the outdoor semantic segmentation challenge on our web page.
Invited SpeakersWe are happy to announce the following invited speakers to our workshop.
The exact program will be announced after the extended abstract & challenge submission deadline.