Abstract | Neglected Tropical Diseases affect more than 24% of the world’s poorest populations. Some NTDs are diagnosed using kato-katz images; the kato-katz technique is the most commonly used diagnostic technique for soil-transmitted helminthiasis. The technique involves microscopic examination of sieved stool samples and tallying the parasitic eggs present. Kato-katz images present a unique and complex problem when it comes to image enhancement for feature detection. The foreground objects of interest (the parasite eggs), have features that match with the rest of the foreground objects in the image, making it difficult to distinguish and differentiate them. Hence sensible noise filtering and feature enhancement strategies are required for successful object identification and diagnosis. This paper will describe the results of studies on noise filtering and image enhancement of kato-katz images. Due to the ‘noisy’ and ‘noise-like’ nature of the images and the large number of objects in the images, typical noise filtering methods produce poor results. To our best knowledge, this is a pioneering work looking at image enhancement and noise filtering of kato-katz images. This work builds upon previous research by the authors concerning kato-katz images |
---|