A Hybrid Deep Learning Approach for Texture Analysis

TL;DR

This paper proposes a hybrid deep learning method combining CNN and SVM for texture analysis, aiming to improve generalization and stability across diverse datasets in applications like remote sensing.

Contribution

It introduces a novel fusion of CNN and SVM to enhance texture classification robustness and generalization across various datasets.

Findings

Improved classification accuracy across multiple datasets

Enhanced stability in texture recognition

Effective combination of CNN features with SVM classifier

Abstract

Texture classification is a problem that has various applications such as remote sensing and forest species recognition. Solutions tend to be custom fit to the dataset used but fails to generalize. The Convolutional Neural Network (CNN) in combination with Support Vector Machine (SVM) form a robust selection between powerful invariant feature extractor and accurate classifier. The fusion of experts provides stability in classification rates among different datasets.