Detecting salient cues through illumination-invariant color ratios

Abstract

This work presents a novel technique for embedding color constancy into a saliency-based system for detecting potential landmarks in outdoor environments. Since multiscale color opponencies are among the ingredients determining saliency, the idea is to make such opponencies directly invariant to illumination variations, rather than enforcing the invariance of colors themselves. The new technique is compared against the alternative approach of preprocessing the images with a color constancy procedure before entering the saliency system. The first procedure used in the experimental comparison is the well-known image conversion to chromaticity space, and the second one is based on successive lighting intensity and illuminant color normalizations. The proposed technique offers significant advantages over the preceding two ones since, at a lower computational cost, it exhibits higher stability in front of illumination variations and even of slight viewpoint changes, resulting in a better correspondence of visual saliency to potential landmark elements.

Introduction

The extraction of reliable visual landmarks for robot localization in outdoor unstructured environments is still an open research problem. One of the main difficulties is that acquired visual information is strongly dependent on lighting geometry (direction and intensity of light source) and illuminant color (spectral power distribution), which change with sun position and atmospheric conditions [29]. In order to overcome these adversities, the acquired images are often submitted to transformations, in an attempt to reduce the dependence on illumination. This desired invariance of color representation to general changes in illumination is called color constancy [1], [6], [21].

In this work, we evaluate three approaches to color constancy applied to a visual landmark detection system. The first two approaches use standard color constancy preprocessing algorithms followed by the landmark detection, while for the third approach, which is the main contribution of this work, we designed a novel color constancy algorithm embedded in the landmark detector.

The paper is organized as follows. In Section 2, the visual saliency and opponent color concepts are introduced, followed by a description of the landmark detection system based on visual saliency. The adopted color model and color constancy techniques used as preprocessing stages are explained in Section 3, together with their connection to the landmark detection system. In Section 4, the proposed visual saliency algorithm, enhanced with embedded color constancy based on color ratios, is described. Finally, in Section 5, all techniques are discussed and compared in the context of saliency-based landmark detection.