Structural similarity metrics for texture analysis and retrieval

Abstract

The development of objective texture similarity metrics for image analysis applications differs from that of traditional image quality metrics because substantial point-by-point deviations are possible for textures that according to human judgment are essentially identical. Thus, structural similarity metrics (SSIM) attempt to incorporate "structural" information in image comparisons. The recently proposed structural texture similarity metric (STSIM) relies entirely on local image statistics. We extend this idea further by including a broader set of local image statistics, basing the selection on metric performance as compared to subjective evaluations. We utilize both intra-and inter-subband correlations, and also incorporate information about the color composition of the textures into the similarity metrics. The performance of the proposed metrics is compared to PSNR, SSIM, and STSIM on the basis of subjective evaluations using a carefully selected set of 50 texture pairs.

A new subjective procedure for evaluation and development of texture similarity metrics

“…In this section we review the structural texture similarity metrics (STSIM) proposed in [6,3]. These metrics compute the similarity of two images x and y by multiplicatively combining a number of terms that compare subband statistics of the two images.…”

“…(4) The STSIM-2 metric proposed in [3] extends the ideas of [6] by including a broader set of local image statistics, motivated by the analysis/synthesis literature [7]. In addition to the terms in (4), STSIM-2 uses terms that compare the crosscorrelation between subbands.…”

“…For example, the Spearman's rank correlation coefficient and Kendall's tau rank correlation coefficient can be used when a relative similarity scale is needed [3], while linear regression can be used when an absolute scale is necessary. In both cases, however, a large number of subjective tests is needed in order to compare the objective and subjective similarity between texture pairs.…”

“…However, such a relationship makes sense only for similar textures. It is difficult even for humans to quantify the similarity of textures that are not similar [3]. Does it make sense to say that sand and tree bark are more dissimilar than marble and snake skin?…”

“…However, our previous work indicates [3,4] that the two attributes should be considered separately. Thus, in the present study we focus only on grayscale textures.…”