Method Details

Details for method 'Scale invariant CNN + CRF'

name	Scale invariant CNN + CRF
challenge	pixel-level semantic labeling
details	We propose an effective technique to address large scale variation in images taken from a moving car by cross-breeding deep learning with stereo reconstruction. Our main contribution is a novel scale selection layer which extracts convolutional features at the scale which matches the corresponding reconstructed depth. The recovered scaleinvariant representation disentangles appearance from scale and frees the pixel-level classifier from the need to learn the laws of the perspective. This results in improved segmentation results due to more effi- cient exploitation of representation capacity and training data. We perform experiments on two challenging stereoscopic datasets (KITTI and Cityscapes) and report competitive class-level IoU performance.
publication	Convolutional Scale Invariance for Semantic Segmentation I. Kreso, D. Causevic, J. Krapac, and S. Segvic GCPR 2016 https://ivankreso.github.io/papers/kreso16gcpr.pdf
project page / code	https://github.com/ivankreso/scale-invariant-cnn
used Cityscapes data	fine annotations, stereo
used external data	ImageNet
runtime	n/a
subsampling	no
submission date	April, 2016
previous submissions