Automatic Detection and Characterization of Parasite Eggs by Image Processing

The accurate identification of parasites allows for the quick diagnosis and treatment of infections. Current state-of-the-art identification techniques require a trained technician to examine prepared specimens by microscope or other molecular methods. In an effort to automate the process and better facilitate the field identification of parasites, approaches are developed to utilize LabVIEW and MATLAB, which are commercially available image processing software packages, for parasite egg identification. The goal of this project is to investigate different image processing techniques and descriptors for the detection and characterization of the following parasite eggs: Ascaris lumbricoides, Taenia sp., and Paragonimus westermani. One manual approach and four automated approaches are used to locate the parasite eggs and gather parasite characterization data. The manual approach uses manual measurements of the parasite eggs within the digital images. The four automated approaches are LabVIEW Vision Assistant scripts, MATLAB separation code, MATLAB cross-section grayscale analysis, and MATLAB edge signature analysis. Forty-four separate measurements were analyzed through the four different approaches. Two types of statistical tests, single factor global Analysis of Variance (ANOVA) test and Multiple Comparison tests, are used to demonstrate that parasite eggs can be differentiated. Thirty-six of the measurements proved to be statistically significant in the differentiation of at least two of the parasite egg types. Of the thirty-six measurements, seven proved to be statistically significant in the differentiation of all three parasite egg types. These results have shown that it is feasible to develop an automated parasite egg detection and identification algorithm through image processing. The automated image processing techniques have proven successful at differentiating parasite eggs from background material. This initial research will be the foundation for future software structure, image processing techniques, and measurements that should be used for automated parasite egg detection.