由粗定位到精提取的图像显著区域检测

图像显著区域检测旨在找出最具信息性的图像,将该任务归纳为一种由粗定位到精提取的处理过程。首先,将图像过分割为超像素,并根据其颜色差异、分形差异及空间分布求得一个表征超像素间相似性的矩阵。依据这个矩阵,利用相似传播算法对超像素聚类;并通过度量类间颜色对比度、类的结构紧凑度与偏离中心度评价每类的显著度。然后,通过比较像素与每类的颜色差异及位置关系更新像素的显著度,最终得到像素精度的、全分辨率的显著性图。对当前流行数据库的实验测试表明,算法具有令人满意的检测效果。

From Rough Location to accurate extraction to Detect Image Salient Region

The task of image salient region detection aims at establishing the most important and informative regions of an image. In this work, we propose a novel method that tackles such task as a process from superpixel-level locating to pixel-level extraction. First, we over-segment the image into superpixels and compute an affinity matrix to estimate the similarity between each two superpixels according to their color contrast, fractal contrast and space distribution. The matrix is applied to aggregate superpixels into several clusters by using affinity propagation (AP) algorithm. To measure the saliency of each cluster, three parameters are taken into account including color contrast, cluster compactness and proximity to the focus. Then, pixel-wise saliency is decided by comparing the color distinction and spatial distance between one pixel and every cluster. We evaluate our algorithm on the publicly available dataset with human annotations, and experimental results show that our approach has competitive performance.