Deep Learning Based Target Tracking and Classification Directly in Compressive Measurement for Low Quality Videos

Abstract

This paper proposes a method based on Multitexton Histogram (MTH) descriptor to identify patterns inimages of human parasite eggs of the following species: Ascaris, Uncinarias, Trichuris, Hymenolepis Nana, Dyphillobothrium-Pacificum, Taenia-Solium, Fasciola Hepatica and Enterobius-Vermicularis. These patterns are represented by textons of irregular shapes in their microscopic images. This proposed method could be used for diagnosis of Parasitic disease and it can be helpful especially in remote places. This paper includes two stages. In the first a feature extraction mechanism integrates the advantages of cooccurrence matrix and histograms to identify irregular morphological structures in the biological images through textons of irregular shape. In the second stage the Support Vector Machine (SVM) is used to classificate the different human parasite eggs. The results were obtaining using a dataset with 2053 human parasite eggs images achieving a success rate of 96,82% in the classification. In addition, this research shows that the proposed method also works with natural images.