自动对焦算法评估线虫脂滴图像的研究

自动对焦是实现线虫自动化筛选的一个重要步骤．在光学显微镜系统中，通过采集同一个视野下不同焦面的图像，再通过清晰度评价函数对这些图像进行运算，得到的最大值被认为是最佳对焦位置．在本研究中，对16种常用的自动对焦算法以及最近提出的一些算法进行了评估，通过评估找出最适合线虫脂滴图像的自动对焦算法，从而搭建一套线虫脂滴自动化筛选系统．同时就对焦精度、运算时间、抗噪声能力、对焦曲线等特征进行了分析评价，结果表明，大多数算法对线虫脂滴图像都有较好的表现，特别是绝对Tenengrad算法在对焦精度上有最好的表现，我们将优选该算法应用到线虫脂滴自动化筛选系统中．

Evaluation of Autofocus Algorithms for Automatic Detection of Caenorhabditis elegans Lipid Droplets

Autofocusing is a fundamental step towards automated microscopic screening of Caenorhabditis elegans. Determining the optimal focus in an optical microscope is based on a clarity-evaluation function that is applied to images acquired from different focuses of the same field. The maximum value of the function is considered as the point of optimal focus. In this paper, 16 autofocus algorithms which were collected from well-known algorithms as well as the most recently proposed focusing algorithms have been evaluated. Through these evaluations, an optimal algorithm was found for C. elegants lipid droplets to set up an automatic screening system. Many features were assessed in this paper, for instance accuracy, computational time, addition of noise, and focusing curve. Our results have shown that most of the algorithms show an overall high performance for this type of image, and absolute Tenengrad algorithm will be our first choice for its best performance considering accuracy.