This article introduces a novel measure called Structure-measure to evaluate non-binary foreground maps in salient object detection. The proposed measure simultaneously evaluates region-aware and object-aware structural similarities between a saliency map (SM) and a ground-truth (GT) map. Traditional measures like AUC, AP, and Fβ often fail to capture structural similarities, which are crucial for accurate salient object detection. The Structure-measure is designed to be efficient and easy to calculate, and it outperforms existing measures on five benchmark datasets. The measure is based on two key characteristics: sharp foreground-background contrast and uniform saliency distribution. The article also presents experiments showing that the Structure-measure performs better than other measures in terms of ranking consistency, robustness to annotation errors, and alignment with human judgments. The results demonstrate that the proposed measure is more effective in evaluating the structural similarity between SM and GT maps, making it a more reliable metric for salient object detection. The article concludes that the Structure-measure provides new insights into evaluating salient object detection models and encourages the saliency community to consider it in future model evaluations and comparisons.This article introduces a novel measure called Structure-measure to evaluate non-binary foreground maps in salient object detection. The proposed measure simultaneously evaluates region-aware and object-aware structural similarities between a saliency map (SM) and a ground-truth (GT) map. Traditional measures like AUC, AP, and Fβ often fail to capture structural similarities, which are crucial for accurate salient object detection. The Structure-measure is designed to be efficient and easy to calculate, and it outperforms existing measures on five benchmark datasets. The measure is based on two key characteristics: sharp foreground-background contrast and uniform saliency distribution. The article also presents experiments showing that the Structure-measure performs better than other measures in terms of ranking consistency, robustness to annotation errors, and alignment with human judgments. The results demonstrate that the proposed measure is more effective in evaluating the structural similarity between SM and GT maps, making it a more reliable metric for salient object detection. The article concludes that the Structure-measure provides new insights into evaluating salient object detection models and encourages the saliency community to consider it in future model evaluations and comparisons.