基于视觉注意机制的彩色图像显著性区域提取

图像显著性区域提取是计算机视觉处理的重要步骤。结合人类视觉心理、生理模型, 提出一种基于视觉注意机制的彩色图像显著性区域提取模型。通过改进的分水岭算法对彩色图像进行预分割, 从而将原图像分成若干子区域, 在此基础上运用提出的区域化空间注意力模型对各个子区域进行显著图计算, 得到最终的显著性区域提取结果。实验结果表明, 提出的显著性区域提取算法可以很好地从彩色图像中得到与视觉注意机制相一致的结果, 且满足实时性要求, 与传统方法相比, 算法提取的区域更完整、更准确。

Saliency detection for color image based on visual attention mechanism

Image saliency detection is an important step for computer vision processing. This paper presented a novel saliency detection algorithm for color image based on visual attention mechanism from human's psychology and physiology. Firstly, it used improved watershed segmentation method to divide the original image into several subdomains. Afterwards, it used regional spatial attention model to compute saliency map for each subdomain to get final saliency detection result. Experiment results show that the presented algorithm can accurately get saliency detection results in short time, and compared with traditional ones, the presented algorithm's results are more intact and more accurate.