• DOI:10.1109/TPAMI.2016.2592911
Corpus ID: 5629671

Tree-Structured Models for Efficient Multi-Cue Scene Labeling

@article{Cordts2017TreeStructuredMF,
  title={Tree-Structured Models for Efficient Multi-Cue Scene Labeling},
  author={Marius Cordts and Timo Rehfeld and M. Enzweiler and Uwe Franke and S. Roth},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year={2017},
  volume={39},
  pages={1444-1454}
}

Marius Cordts, Timo Rehfeld, +2 authors S. Roth

Published 2017

Computer Science, Medicine

IEEE Transactions on Pattern Analysis and Machine Intelligence

We propose a novel approach to semantic scene labeling in urban scenarios, which aims to combine excellent recognition performance with highest levels of computational efficiency. To that end, we exploit efficient tree-structured models on two levels: pixels and superpixels. At the pixel level, we propose to unify pixel labeling and the extraction of semantic texton features within a single architecture, so-called encode-and-classify trees. At the superpixel level, we put forward a multi-cue… CONTINUE READING

View on IEEE

doc-08-1g-docs.googleusercontent.com

Save to Library

Create Alert

Cite

Launch Research Feed