Autofocusing software issues in automated fluorescence microscopy
| Authors | |
|---|---|
| Year of publication | 2009 |
| Type | Conference abstract |
| MU Faculty or unit | |
| Citation | |
| Description | This work aims at presenting the basic issues with autofocusing software for automated fluorescence microscopy, together with possible solutions. Namely, three key issues and two techniques are discussed, and the difference between confocal and wide-field microscopy is covered in each case. There are two key factors in image autofocusing: the focus function and the autofocusing algorithm. The algorithm steps along the z-axis and evaluates focus values of each acquired image according to the focus function, in order to find its maximum. In reality, the objects often do not lie in one plane, which results in multimodality of the focus function. Thus all the significant local maxima of this function should be found. Moreover, the peaks can sometimes be coalesced together, especially if they are close to each other, or if one of them is too dominant. This contribution suggests solutions to the following: 1. Selection of suitable focus function (this is especially difficult in wide-field mode, due to strong influence of the out-of-focus glare). 2. Discovering those hidden interesting planes of focus. 3. Efficiency of the autofocusing algorithms. The first two items can be solved using a new autofocusing technique, which consists in dividing the acquired image into several parts, and treating each part differently. Submissive planes of focus can become more dominant when only subpart of the image is considered. Also the undesirable influence of the out-of-focus glare can be partially eliminated. The efficiency of the algorithm can be increased by adaptively varying step size. This technique can be combined with the previous one in the case of confocal laser scanning microscope system. Concluding remarks are devoted to the unsolved issues. |
| Related projects: |