DIFFERENTIAL INVARIANTS FOR SE(2)-EQUIVARIANT NETWORKS - Archive ouverte HAL Access content directly
Conference Papers Year :

DIFFERENTIAL INVARIANTS FOR SE(2)-EQUIVARIANT NETWORKS

Abstract

Symmetry is present in many tasks in computer vision, where the same class of objects can appear transformed, e.g. rotated due to different camera orientations, or scaled due to perspective. The knowledge of such symmetries in data coupled with equivariance of neural networks can improve their generalization to new samples. Differential invariants are equivariant operators computed from the partial derivatives of a function. In this paper we use differential invariants to define equivariant operators that form the layers of an equivariant neural network. Specifically, we derive invariants of the Special Euclidean Group SE(2), composed of rotations and translations, and apply them to construct a SE(2)-equivariant network, called SE(2) Differential Invariants Network (SE2DINNet). The network is subsequently tested in classification tasks which require a degree of equivariance or invariance to rotations. The results compare positively with the state-of-the-art, even though the proposed SE2DINNet has far less parameters than the compared models.
Fichier principal
Vignette du fichier
main.pdf (252.02 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03703287 , version 1 (23-06-2022)
hal-03703287 , version 2 (18-07-2022)

Identifiers

Cite

Mateus Sangalli, Samy Blusseau, Santiago Velasco-Forero, Jesus Angulo. DIFFERENTIAL INVARIANTS FOR SE(2)-EQUIVARIANT NETWORKS. 29th IEEE International Conference on Image Processing (IEEE ICIP), Oct 2022, Bordeaux, France. ⟨hal-03703287v2⟩
90 View
77 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More