Convolutional Neural Networks Rarely Learn Shape for Semantic Segmentation.

Yixin Zhang Maciej A. Mazurowski

Published in: CoRR (2023)

Keyphrases

semantic segmentation
convolutional neural networks
superpixels
label transfer
conditional random fields
weakly supervised
object categories
shape model
scene classification
object segmentation
pascal voc
pairwise
object classes
object class