Automatic Identification And Classification Of Parasite Eggs In Microscopic Fecal Images

In view of the great harm to human health caused by parasitic diseases,the diagnosis and treatment of parasitic diseases has become a very important part of clinical medicine.The diagnosis of parasitic diseases in clinical medicine is realized by detecting whether there are parasitic eggs in human feces.The traditional inspection method is carried out by the medical staff,which has the disadvantages of low detection efficiency,large workload and low detection accuracy.With the development of image processing technology and pattern recognition technology,the identification and classification of parasitic eggs were gradually converted to computer processing,this method not only can improve the detection efficiency and accuracy,but also can reduce the workload of medical inspectors.The automatic identification and classification technology of parasite eggs in microscopic fecal medical images has wide application prospect and wide development space.The traditional fecal sample image acquisition using high power microscope(40x),a small range of image acquisition,the number of images needed to be collected too much,resulting in the identification of parasitic eggs more slowly.Therefore,the fecal samples collected in this study were collected using a low magnification(10x)microscope.In this paper,the automatic identification and classification of 8 kinds of parasitic eggs,which are commonly used in fecal samples collected under low power microscope.The main contents of this paper are as follows.Firstly,the microscopic medical images were pretreated.The parasite eggs in the image of the stool samples collected under the low magnification microscope were small and the images contained large amounts of impurities.By comparing the results of several image segmentation techniques,the method of image segmentation based on mathematical morphology and feature selection is used in this paper.Secondly,the feature selection and feature extraction of parasitic eggs were studied.Based on the observation and analysis of parasite eggs,geometric shapes of parasite eggs can be used as the basis for identification and classification of parasite eggs,at the same time,the distribution of parasite egg contents have structural texture features of parasite eggs can also be used as the basis for recognition and classification.Finally,the identification and classification of parasitic eggs were studied.The characteristics of different types of eggs are different,among which the characteristics of the eggs of the liver and the insects are different from those of the other six parasites,which are identified by the feature screening method.The remaining parasitic 6 eggs are similar in character,and this paper uses the improved KNN classifier to classify them.The characteristics of this study are that the fecal images collected at low magnification microscopic microscope are complex and have an effective identification and classification of up to eight common parasite eggs.The experimental results show that the correct recognition rate of the eight eggs is more than 90%,and the accurate recognition rate of the fasciolopsis eggs is 98%,which meets the requirements of clinical examination.At present,the technology studied in this paper has been put into use in some domestic hospitals.