MorphoActivation: Generalizing ReLU activation function by mathematical morphology
Machine Learning
2022-07-15 v1 Discrete Mathematics
Image and Video Processing
Signal Processing
Applications
Abstract
This paper analyses both nonlinear activation functions and spatial max-pooling for Deep Convolutional Neural Networks (DCNNs) by means of the algebraic basis of mathematical morphology. Additionally, a general family of activation functions is proposed by considering both max-pooling and nonlinear operators in the context of morphological representations. Experimental section validates the goodness of our approach on classical benchmarks for supervised learning by DCNN.
Keywords
Cite
@article{arxiv.2207.06413,
title = {MorphoActivation: Generalizing ReLU activation function by mathematical morphology},
author = {Santiago Velasco-Forero and Jesús Angulo},
journal= {arXiv preprint arXiv:2207.06413},
year = {2022}
}