A cell is the structural and functional unit of all known living organisms, and its three-dimensional shape is an interesting
research topic and having many applications in biology. Usually, cells are kept surrounded with some liquid materials on
glass plates. In obtained image sequence, liquid material causes unwanted background in the images, and some virtual
images due to the glass plates occurs, which makes difficulty to recover the three-dimensional shape of the cell.
Therefore, conventional optical passive methods for three-dimensional shape recovery do not compute depth map
accurately. The purpose of this work is to reconstruct three-dimensional shape of HeLa cell by applying shape from
focus (SFF) method. SFF method is one of the optical passive methods to estimate three-dimensional shape by using
focal information from image sequence. To overcome problems from transparency and reflection, transparent part is
segmented from images by using the fact that background of the cell does not have focal point, and an original image
sequence is divided into two image sequences for real and virtual part by finding two focused points in itself. For more
accurate segmentation of the background part, the labeling method is used, and for automatically dividing an original
image sequence into two image sequences, the iterative threshold selection method is used. The proposed approach is
tested by using HeLa cell which is one of the most famous cells in biological research area. The experimental result
demonstrates the effectiveness.
This paper presents the use of Genetic Algorithm as a search method for focus measure in Shape From Focus (SFF). Previous methods compute focus value for each pixel locally by summing all values within a small window. This summation is a good approximation of focus quality, but is not optimal one. The Genetic Algorithm is used
as a fine tuning process in which a measure of best focus is used as the fitness function corresponding to motion parameter values which make up each gene. The experimental results show that the proposed method performs better than previous algorithms such as Sum of the Modified Laplacian(SML), Grey Level Variance(GLV) and
Tenenbaum Focus Measure. The results are compared using root mean square error(RMSE) and correlation. The experiments are conducted using objects simulated cone, real cone and TFT-LCD color filter<sup>1</sup> to evaluate performance of the proposed algorithm.