Empirical curvelet based Fully Convolutional Network for supervised texture image segmentation

TL;DR

A method to build a unique empirical curvelet filter bank adapted to a given dictionary of textures is proposed and it is shown that the output of these filters can be used to build efficient texture descriptors utilized to finally feed deep learning networks.

Abstract

In this paper, we propose a new approach to perform supervised texture classification/segmentation. The proposed idea is to feed a Fully Convolutional Network with specific texture descriptors. These texture features are extracted from images by using an empirical curvelet transform. We propose a method to build a unique empirical curvelet filter bank adapted to a given dictionary of textures. We then show that the output of these filters can be used to build efficient texture descriptors utilized to finally feed deep learning networks. Our approach is finally evaluated on several datasets and compare the results to various state-of-the-art algorithms and show that the proposed method dramatically outperform all existing ones.

Figures (12)

• Figure 1: Magnitude of the Fourier spectrum segmentation and empirical wavelet construction principle.
• Figure 2: Examples of angular boundaries, i.e. the sets $\Omega_{\theta}^{i}$ (the vertical red lines), obtained from different input textured images.
• Figure 3: Proposed texture feature extraction process.
• Figure 4: Standard fully convolutional network makes dense predictions for per-pixel semantic segmentation
• Figure 5: Network architecture used to segment grayscale images.