基于颜色通道比较的显著性检测

为了快速有效地检测图像中的显著性区域，利用颜色通道差异信息和空间信息，提出了一种基于颜色通道比较的显著性检测算法. 该算法将显著性检测任务看作对图像前景区域和背景区域进行分离和识别的过程，首先基于RGB颜色空间构造了一组反映不同色彩信息的颜色通道，通过通道比较将不同颜色的区域分离开，然后通过分析空间信息选择出属于前景的区域，最后对所有前景区域进行合并得到最终的显著图. 我们在两个公开数据集上与现有的一些检测算法进行了对比试验，结果表明该算法具有较低的计算复杂度以及较高的检测准确率和召回率，证明该算法是简单有效的.

Saliency Detection Based on Comparison of Color Channels

To detect the salient regions in images quickly and efficiently, using information of difference between color channels and spatial information, this paper proposes a saliency detection approach based on comparison of color channels. The algorithm regarded the saliency detection task as a process of separating and identifying the foreground region and the background region of the image. Firstly, it defines a set of color channels based with RGB color space, and separates regions with different colors in image by making comparisons between different channels. Then, it selects foreground regions by analyzing spatial information. Lastly, it combines all foreground regions to obtain the final saliency map. We evaluate and compare the performances of our algorithm against previous algorithms on two public datasets. The experimental results show that our algorithm has lower computational complexity and high precision and recall rate, and prove that our algorithm is simple and effective.