School of Communication and Information Engineering, Shanghai University, Shanghai 200072, China.
Opt Lett. 2013 Mar 1;38(5):700-2. doi: 10.1364/OL.38.000700.
We propose an efficient regional histogram (RH)-based computation model for saliency detection in natural images. First, the global histogram is constructed by performing an adaptive color quantization on the original image. Then multiple RHs are built on the basis of the region segmentation result, and the color-spatial similarity between each pixel and each RH is calculated accordingly. Two efficient measures, distinctiveness and compactness of each RH, are evaluated based on the color difference with the global histogram and the color distribution over the whole image, respectively. Finally, the pixel-level saliency map is generated by integrating the color-spatial similarity measures with the distinctiveness and compactness measures. Experimental results on a dataset containing 1000 test images with ground truths demonstrate that the proposed saliency model consistently outperforms state-of-the-art saliency models.
我们提出了一种基于高效区域直方图（RH）的自然图像显著度检测计算模型。首先，通过对原始图像进行自适应颜色量化来构建全局直方图。然后，基于区域分割结果构建多个 RH，并相应地计算每个像素与每个 RH 之间的颜色-空间相似度。基于与全局直方图的颜色差异和整个图像上的颜色分布，分别评估每个 RH 的两个有效度量值，即独特性和紧凑性。最后，通过将颜色-空间相似度度量与独特性和紧凑性度量相结合，生成像素级显著图。在包含 1000 张测试图像和地面实况的数据集上的实验结果表明，所提出的显著度模型始终优于最先进的显著度模型。
