Image Segmentation under Microscope Based on Contour Feature Extraction

  • Rui Li
Keywords: Image Segmentation, Microscope, Feature Extraction, OTSU Algorithm, Adaptive Genetic Algorithm

Abstract

Contour feature extraction is a key step in microorganism image analysis because image segmentation, object
isolation, feature extraction and parameter measurement will convert the original microorganism image into a
more abstract and compact form, making possible higher-level analysis and understanding.Conventional 1D
OTSU does not lead to a stable segmentation result for the general microorganism images, or even those images
with an unconspicuous two-peak histogram while 2D OTSU involves large computation and takes much time.
This paper combines the adaptive genetic algorithm (GA) with OTSU threshold segmentation method. It
guarantees image segmentation accuracy by using that OTSU needs to take into consideration the relevant
information in neighborhood space, compensates for the deficiencies of OTSU threshold method with the
globalstrengths of GA, and introduces adaptive factors to further improve the accuracy, operating rate, and
contour feature extraction efficiency of microorganism image and ensure the stability and accuracy to find the
optimal segmentation threshold. The experiment result shows that this algorithm works feasibly and reliably,
achieves ideal segmentation results, and implements faster and more stable image segmentation than traditional
methods.

Published
2020-04-01