From CNNs to Shift-Invariant Twin Models Based on Complex Wavelets - Archive ouverte HAL Access content directly
Preprints, Working Papers, ... Year : 2024

From CNNs to Shift-Invariant Twin Models Based on Complex Wavelets

Abstract

We propose a novel method to increase shift invariance and prediction accuracy in convolutional neural networks. Specifically, we replace the first-layer combination "real-valued convolutions → max pooling" (RMax) by "complex-valued convolutions → modulus" (CMod), which is stable to translations, or shifts. To justify our approach, we claim that CMod and RMax produce comparable outputs when the convolution kernel is band-pass and oriented (Gabor-like filter). In this context, CMod can therefore be considered as a stable alternative to RMax. To enforce this property, we constrain the convolution kernels to adopt such a Gabor-like structure. The corresponding architecture is called mathematical twin, because it employs a well-defined mathematical operator to mimic the behavior of the original, freely-trained model. Our approach achieves superior accuracy on ImageNet and CIFAR-10 classification tasks, compared to prior methods based on low-pass filtering. Arguably, our approach's emphasis on retaining high-frequency details contributes to a better balance between shift invariance and information preservation, resulting in improved performance. Furthermore, it has a lower computational cost and memory footprint than concurrent work, making it a promising solution for practical implementation.
Fichier principal
Vignette du fichier
preprint.pdf (2.31 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-03880520 , version 1 (01-12-2022)
hal-03880520 , version 2 (21-04-2023)
hal-03880520 , version 3 (31-05-2024)

Licence

Identifiers

  • HAL Id : hal-03880520 , version 3

Cite

Hubert Leterme, Kévin Polisano, Valérie Perrier, Karteek Alahari. From CNNs to Shift-Invariant Twin Models Based on Complex Wavelets. 2024. ⟨hal-03880520v3⟩
111 View
53 Download

Share

Gmail Mastodon Facebook X LinkedIn More