Phase congruence measurement for image similarity assessment

Abstract

In the performance assessment of an image processing algorithm, an image is often compared with an available reference. Measuring image similarity can be achieved in many ways; comparison algorithm varies from pixel-based mean square error method to structure-based image quality index. In this paper, we present a new feature-based approach that utilizes image phase congruency measurement to quantify the assessment of the similarities or differences between two images. Test results with standard images and industrial inspection images are presented.

Introduction

The assessment of an image can be carried out by comparing it with a reference image, which is assumed to be perfect for a particular application. Usually, such comparison is implemented on pixel-based operations, like mean square error (MSE) or root mean square error (RMSE). However, such operations’ performance is questionable because the same MSE or RMSE value does not always assure a comparable image similarity under different distortion to perceptually significant features (Paolo, 1998).

According to Wilson et al. (1997), there are three major methods for comparing images: human perception, objective measures based on theoretical models, and subjective measures defined mathematically. In their publication (Wilson et al., 1997), they used a distance measure of two sets of pixels to characterize the numerical difference between two images. Metrics for binary image (Δ_b) and gray-scale image (Δ_g) comparison were developed. Lees and Henshaw (1986) used a phase-only approach for printed circuit board (PCB) inspection. The phase-only imaging has the advantages of being light intensity invariant, insensitive to illumination gradients, and tolerant to misregistration. Lorenzetoo and Kovesi proposed to compute phase difference between images using Gabor filters (Leronzetto and Kovesi, 1999). One strength of this algorithm resides in its ability to discriminate simple translation from distortion.

In practical applications, some post-processing operations largely depends on the availability of image features. Operations, like classification, segmentation, and quantification, are often carried out in a feature space. Therefore, the availability of image features plays an important role for further analysis. This paper proposes a new feature-based method to quantitatively assess image similarity by employing the phase congruency measurement suggested by Kovesi, 2000, Kovesi, 1996. Phase congruency provides an absolute measure of image features such as step edges, lines and Mach bands; it is viewing condition-independent and invariant to changes in illumination and magnification. A local cross-correlation of the phase congruence map can then be calculated between a processed image and a reference image. The averaged cross-correlation value provides a quantitative assessment of the overall image similarity. Experiments are carried out on some standard images and on nondestructive inspection images.

The rest of the paper is organized as follows: in Section 2, a brief review of current available techniques is presented. Then, the proposed feature-based metric is described. Experimental results are demonstrated in Section 3. Finally Section 4 is a conclusion.