Supplementary Material for paper1035
DepthCut: Accurate Depth Edge EstimationUsing Multiple Unreliable Channels
--------------------------------------------------------------------------

This zip file contains:

-- inputchannels: all the input channels for the camera dataset.

-- groundtruth: our manually generated ground truth depth edges for the camera dataset.

-- depthedges: the depth edges for our method with each of the 3 disparity methods. Methods can be identified following suffixes:
    _training165_epoch1000_edges : all channels (our method) with mc-cnn disparities
    _training172_epoch1000_edges : all channels (our method) with dispnet disparities
    _training171_epoch1000_edges : all channels (our method) with mono disparities

-- depthedges_globalized: globalized depth after applying spectral globalization from Arbelaez et al. [1]. Same suffixes apply.

-- ucm: ultrametric contour map after applying the UCM framework from Arbelaez et al. [1] and performing the boundary smoothness enhancement. Same suffixes apply.

-- ucm_vis: visualization of the same ultrametric contour map. Same suffixes apply.


[1] P. Arbelaez, M. Maire, C. Fowlkes and J. Malik, "Contour Detection and Hierarchical Image Segmentation," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 33, no. 5, pp. 898-916, May 2011.
doi: 10.1109/TPAMI.2010.161