Lymphatic Filariasis detection in microscopic images

Abstract

In Africa, the propagation of parasites like the lymphatic filariasis is complicating seriously the efforts of health professionals to cure certain diseases. Although there are medicines capable to treat the lymphatic filariasis, this needs to be discovered firstly which is not always an easy task having into account that in most countries affected by this disease it can only be detected at night (nocturne). The lymphatic filariasis is then, a parasitical infection which can originate changes or ruptures in the lymphatic system as well as an abnormal growth of certain areas of the body causing pain, incapacity and social stigma. Approximately 1.23 billion people in 58 countries from all over the world are threatened by this disease which requires a preventive treatment to stop its propagation which makes it even more important for the existence of a mechanism that is less costly and more agile in the analysis of a blood smear to verify the existence of microfilariae (little worms that are produced by other adult worms while housed in the lymphatic system). The lymphatic filariasis is caused by an infection with nematodes ("roundworms") of the Filariodidea family in which three types are inserted: Wuchereria Bancroft, responsible for 90% of all cases; Brugia Malayi, responsible for almost every remaining; B.Timori also causing the disease. All three have characteristics that can differentiate them which allow them to be identified. The current identification process of the disease consists on the analysis of microfilariae in a blood smear with a blood sample through a microscope and its identification by the observer. Taking this into account, it is intended to develop image analysis and processing techniques for the recognition and counting of the two principal types of filarial worms from a thin blood smear, a smartphone and a portable microscope making the detection possible without the need of a health professional and consequent automation of the process. To make this possible an adapter smartphone-microscope can be used to obtain an image with the magnification of 40x3. The images can then be analyzed in a server or in the smartphone, if it has enough processing for it. It is expected from this process that the need to resort to labs to process the blood smear gets fulfilled making the process more accessible and agile instead of costly and slow. For the detection of the parasites from the acquired images it is intended to implement, experiment and choose the more adequate operations. These comprise pre-processing operationswith the goal to enhance the acquired images and eliminate possible artifacts prevenient from the acquisition system. However, the principal operations should be those that allow the verification of existence or nonexistence, recognition and classification of the pretended parasites. Processing and analysis techniques that are common in these processes are based in the extraction of features (e.g. SIRF, SURF, and FLANN) template similarity, edge detection and description of contours and recognition of statistical patterns. Once detected and recognized one or more parasites and its types should be defined and used a rule to declare the presence of the disease and its stage.