OAGM is the Austrian Society for pattern recognition. The paper deals with the automatic segmentation of buildings out of images – exactly to the pixel-level. This is quite difficult because buildings are often displayed in front of complex backgrounds or are sometimes hidden by objects such as cars or trees.
The main contribution of the presented work is a neural network that is able to segment buildings fully automatically out of images as well as a method for post-processing that refines the building boundaries and thereby further improves the segmentation.
The paper won the best poster award at the workshop. Congratulations to Muntaha Sekeena and Matthias Zeppelzauer!