Guided Image Filter and SVM-Based Automated Classification of Microscopy Images

Abstract

Microscopy images are acquired by capturing the microscopic view of blood sample under a microscope using a camera. The image quality is not that reliable further making bacteria classification a tedious task. Further manual classification requires a specialist recommendation. Guided image filter (GIF) has been a suitable filter for contrast enhancement of the images. Otsu thresholding (OT) has been a suitable algorithm for segmentation of microscopy image. Scale invariant feature transform has been an appropriate method for feature extraction. Support vector machines (SVMs) classifier is a suitable classifier with a large dataset. In this paper, a combinative approach of all the aforesaid methods is proposed on the bacterial microscopy images for the classification of the bacterial cells in the microscopy images. The image quality assessment (IQA) of the enhanced image is evaluated using parameters like standard deviation (SD) and entropy. The performance evaluation of the classifier has been carried out using confusion matrix.