Unsupervised Parallel Extraction based Texture for Efficient Image Representation

TL;DR

This paper introduces an unsupervised texture feature extraction method using Concurrent Self-Organizing Maps (CSOM) for more effective image representation, particularly in medical image analysis.

Contribution

It proposes a novel CSOM-based feature extraction approach that outperforms traditional single SOM methods in representing image content.

Findings

CSOM features better represent image content than single SOM features.

Improved decision accuracy in Computer-Aided Diagnosis (CAD) systems.

Validated on MIAS mammographic dataset.

Abstract

SOM is a type of unsupervised learning where the goal is to discover some underlying structure of the data. In this paper, a new extraction method based on the main idea of Concurrent Self-Organizing Maps (CSOM), representing a winner-takes-all collection of small SOM networks is proposed. Each SOM of the system is trained individually to provide best results for one class only. The experiments confirm that the proposed features based CSOM is capable to represent image content better than extracted features based on a single big SOM and these proposed features improve the final decision of the CAD. Experiments held on Mammographic Image Analysis Society (MIAS) dataset.