为得到与主观评价一致的衡量方法, 结合视觉注意机制与边缘展宽衡量, 提出一种显微图像清晰度评价指标.根据不同空间位置的像素灰度差距特征, 计算视觉注意索引图来模拟人眼视觉系统观察景物的特性, 从而赋予不同区域与像素不同的权重, 通过分割实现对潜在对焦区域的提取.利用边缘展宽方法衡量这些区域, 以水平与竖直方向的平均边缘展宽作为最后的清晰度指标.设计主观实验与客观实验, 并与其他客观方法对比, 结果表明该方法能有效克服现有方法的缺陷, 解决显微图像清晰度评价中的关键问题, 实现与主观评价一致的显微图像客观评价.

To achieve the consistency with subjective evaluation, the paper proposes a microscopy image sharpness metric using visual attention mechanism and edge spreading measurement. According to the gray difference of pixels at different spatial location, visual attention index map is calculated to simulate the way of human vision system observing the features of scenery. By this way, weight is given for different regions and pixels, and the potential focuing areas is extract by segmentation. Using edge spreading evaluation in this areas, an average edge spreading of horizontal and vertical direction is obtained, which is considered as the last sharpness metric. The authors design experiments for subjective assessment and objective evaluation. The experiment and comparison demonstrate that, the proposed method can effectively solve the key problems in the evaluation of microscopy image, and satisfy the consistency of objective and subjective evaluation.