Approximation of dilation-based spatial relations to add structural constraints in neural networks

TL;DR

This paper introduces a fast, convolution-based approximation of dilation for modeling spatial relations in neural networks, enhancing training efficiency while maintaining acceptable accuracy.

Contribution

It proposes a novel, computationally efficient approximation of dilation using convolutions, suitable for neural network training with structural spatial constraints.

Findings

The approximation is faster than previous methods.

It maintains comparable accuracy in modeling spatial relations.

Applicable to neural networks requiring structural regularization.

Abstract

Spatial relations between objects in an image have proved useful for structural object recognition. Structural constraints can act as regularization in neural network training, improving generalization capability with small datasets. Several relations can be modeled as a morphological dilation of a reference object with a structuring element representing the semantics of the relation, from which the degree of satisfaction of the relation between another object and the reference object can be derived. However, dilation is not differentiable, requiring an approximation to be used in the context of gradient-descent training of a network. We propose to approximate dilations using convolutions based on a kernel equal to the structuring element. We show that the proposed approximation, even if slightly less accurate than previous approximations, is definitely faster to compute and therefore more suitable for computationally intensive neural network applications.