A Consolidated Approach to Convolutional Neural Networks and the Kolmogorov Complexity

TL;DR

This paper explores using Kolmogorov Complexity-based normalized compression distance within Convolutional Neural Networks to improve unsupervised classification of retinal cell images for age-related macular degeneration therapy.

Contribution

It introduces an innovative algorithmic similarity measure based on Kolmogorov Complexity and integrates it into CNNs for unsupervised cellular image classification.

Findings

Demonstrates the feasibility of Kolmogorov Complexity in image similarity measurement.

Shows potential for improved unsupervised classification accuracy.

Provides a foundation for automated cell classification in medical applications.

Abstract

The ability to precisely quantify similarity between various entities has been a fundamental complication in various problem spaces specifically in the classification of cellular images. Contemporary similarity measures applied in the domain of image processing proposed by the scientific community are mainly pursued in supervised settings. In this work, we will explore the innovative algorithmic normalized compression distance metric based on the information theoretic concept of Kolmogorov Complexity. Additionally we will observe its possible implementation in Convolutional Neural Networks to facilitate and automate the classification of Retinal Pigment Epithelial cell cultures for use in Age Related Macular Degeneration Stem Cell therapy in an unsupervised setting.