At the CVPR (Computer Vision and Pattern Recognition) conference from June 14-19, 2020, latest findings in the field of computer vision and pattern recognition will be discussed. This powerful forum brings together leading academics and researchers from industry to promote the understanding, application and further development of computer vision, AI and machine learning on a global scale.
Matthias Rottmann from the Bergische Universität Wuppertal as well as Philipp Oberdiek and Gernot A. Fink, both from the Technical University of Dortmund, will participate in the SAIAD (Safe Artificial Intelligence for Automated Driving) workshop with a paper. The paper will present a novel pipeline for semantic segmentation that uses deep neural network prediction to identify out-of-distribution (OOD) segments. After feature extraction and dimensionality reduction a retrieval analysis is presented to evaluate the relevance of the found OOD concepts. More information can be found here: Publications.