@InProceedings{birmingham-muscat-belz:2018:W18-65,
  author    = {Birmingham, Brandon  and  Muscat, Adrian  and  Belz, Anja},
  title     = {Adding the Third Dimension to Spatial Relation Detection in 2D Images},
  booktitle = {Proceedings of the 11th International Conference on Natural Language Generation},
  month     = {September},
  year      = {2018},
  address   = {Tilburg University, The Netherlands},
  publisher = {Association for Computational Linguistics},
  pages     = {146--151},
  abstract  = {Spatial relation detection in images has become a popular subject in image description research recently. A range of different language and geometric features have been used in this context, but methods have not so far used explicit information about the third dimension (depth), except when manually added to annotations. The lack of such information hampers detection of many different spatial relations that are inherently 3D. In this paper, we use a fully automatic method for creating a depth map of an image and derive several different object-level depth features from it which we add to an existing feature set to test the effect on spatial relation detection. We show that performance increases are obtained by adding depth features in all scenarios tested.},
  url       = {http://www.aclweb.org/anthology/W18-6517}
}