MorphPool: Efficient Non-linear Pooling & Unpooling in CNNs

Open Access
Authors
Publication date 2022
Book title The 33rd British Machine Vision Conference Proceedings
Book subtitle BMVC 2022 : 21st-24th November 2022, London, UK
Event 33rd British Machine Vision Conference
Article number 56
Number of pages 14
Publisher BMVA Press
Organisations
  • Faculty of Science (FNWI) - Informatics Institute (IVI)
Abstract
Pooling is essentially an operation from the field of Mathematical Morphology, with max pooling as a limited special case. The more general setting of MorphPooling greatly extends the tool set for building neural networks. In addition to pooling operations, encoder-decoder networks used for pixel-level predictions also require unpooling. It is common to combine unpooling with convolution or deconvolution for up-sampling. However, using its morphological properties, unpooling can be generalised and improved. Extensive experimentation on two tasks and three large-scale datasets shows that morphological pooling and unpooling lead to improved predictive performance at much reduced parameter counts.
Document type Conference contribution
Note With supplementary material and video
Language English
Published at https://doi.org/10.48550/arXiv.2211.14037
Published at https://bmvc2022.mpi-inf.mpg.de/56/
Other links https://github.com/rickgroen/morphpool
Downloads
0056 (Final published version)
Supplementary materials
Permalink to this page
Back