Comparison of Filtering Techniques for Cell Image Segmentation

Abstract

In segmentation of cell image for the objective diagnosis by medical
laboratory experts, there are some preprocessing activities that is
carried out before segmentation. The preprocessing includes noise
removal. This paper proposes the comparison of filtering techniques
for cell image segmentation in order to state the best filter suitable
for noise removal. The filter being compared are mean, median and
Gaussian filters. MATLAB IDE software was used to implement each
of the filter’s algorithm. The results obtained were evaluated and
compared to identify which of the filter is best used for the
preprocessing of the cell images. The result shows that the median
filter outperformed the mean and Gaussian filters with 92.70% of
accuracy for the median filter.