p-adic Cellular Neural Networks: Applications to Image Processing

TL;DR

This paper introduces two novel p-adic cellular neural networks for image processing, demonstrating their ability to detect edges and reduce noise while preserving image features through stability analysis and numerical simulations.

Contribution

The work presents new p-adic CNN models for image analysis, including hierarchical edge detection and noise reduction, with a focus on their stability and practical implementation.

Findings

Networks effectively detect edges in grayscale images.

Networks can reduce Gaussian noise while preserving edges.

Numerical simulations confirm practical applicability.

Abstract

The p-adic cellular neural networks (CNNs) are mathematical generalizations of the neural networks introduced by Chua and Yang in the 80s. In this work we present two new types of CNNs that can perform computations with real data, and whose dynamics can be understood almost completely. The first type of networks are edge detectors for grayscale images. The stationary states of these networks are organized hierarchically in a lattice structure. The dynamics of any of these networks consists of transitions toward some minimal state in the lattice. The second type is a new class of reaction-diffusion networks. We investigate the stability of these networks and show that they can be used as filters to reduce noise, preserving the edges, in grayscale images polluted with additive Gaussian noise. The networks introduced here were found experimentally. They are abstract evolution equations on spaces of real-valued functions defined in the p-adic unit ball for some prime number p. In practical applications the prime p is determined by the size of image, and thus, only small primes are used. We provide several numerical simulations showing how these networks work.