Structure-Texture Decomposition of RGB-D Images

Abstract

In this paper, we study the problem of separating texture from structure in RGB-D images. Our structure preserving image smoothing operator is based on the region covariance smoothing (RCS) method in [16] that we present a number of modifications to this framework to make it depth-aware and increase its effectiveness. In particular, we propose to incorporate three geometric depth features, namely height above ground, angle with gravity and horizontal disparity to the pool of image features used in that study. We also suggest to use a new kernel function based on KL-divergence between the distributions of extracted features. We demonstrate our approach on challenges images from NYU-Depth v2 Dataset [24], achieving more accurate decompositions than the state-of-the-art approaches which do not utilize any depth information.

Authors and Affiliations

Aykut Erdem*| Hacettepe University, Department of Computer Engineering, Ankara, Turkey.

Keywords

Related Articles

Classification of Siirt and Long Type Pistachios (Pistacia vera L.) by Artificial Neural Networks

Quality is one of the important factors in agricultural products marketing. Grading machines have great role in quality control systems. The most efficient method used in grading machines today is image processing. This...

Improving Intrusion Detection using Genetic Linear Discriminant Analysis

The objective of this research is to propose an efficient soft computing approach with high detection rates and low false alarms while maintaining low cost and shorter detection time for intrusion detection. Our results...

A Bee Colony Optimization-based Approach for Binary Optimization

The bee colony optimization (BCO) algorithm, one of the swarm intelligence algorithms, is a population based iterative search algorithm. Being inspired by collective bee intelligence, BCO has been proposed for solving di...

PID Parameters Prediction Using Neural Network for A Linear Quarter Car Suspension Control

Providing control for suspension systems in vehicles is an enhancing factor for comfort and safety. With the improvement of control conditions, it is possible to design a cost-efficient controller which will maintain opt...

A Note on Entropy Subsethood Relationship

We comment on subsethood measure defined by Kosko and Young and give some new aspects of these measures. Finally we would like to discard the entropy subsethood relationship established by the authirs. We present some pr...

Download PDF file
  • EP ID EP808
  • DOI 10.18201/ijisae.2016426381
  • Views 387
  • Downloads 23

How To Cite

Aykut Erdem* (2016). Structure-Texture Decomposition of RGB-D Images. International Journal of Intelligent Systems and Applications in Engineering, 4(4), 111-118. https://europub.co.uk/articles/-A-808