Unsupervised identification of malaria parasites using computer vision

Abstract

Malaria in human is a serious and fatal tropical disease. This disease results from Anopheles mosquitoes that are infected by Plasmodium species. The clinical diagnosis of malaria based on the history, symptoms and clinical findings must always be confirmed by laboratory diagnosis. Laboratory diagnosis of malaria involves identification of malaria parasite or its antigen / products in the blood of the patient. Manual diagnosis of malaria parasite by the pathologists has proven to become cumbersome. Therefore, there is a need of automatic, efficient and accurate identification of malaria parasite.In this paper, we proposed a computer vision based approach to identify the malaria parasite from light microscopy images. This research deals with the challenges involved in the automatic detection of malaria parasite tissues. Our proposed method is based on the pixel based approach. We used K-means clustering (unsupervised approach) for the segmentation to identify malaria parasite tissues.

Malaria Parasitaemia Estimation � A new perspective (Preprint)

“…Nasir et al [48], used the saturation channel of image in HSI colour space and moving Knearest neighbour (MKNN) to differentiate parasite containing red blood cell. Khan et al [49] performed clustering on the 'b' component of the image converted to Lab colour space for obtaining parasite region.…”

Segmentation of Plasmodium vivax phase on digital microscopic images of thin blood films using colour channel combination and Otsu method

“…Khan et al [6] developed a segmentation method using b* channel from the colour space of L*a*b* and KMeans to obtain the tissue of P. vivax. Nasir et al [3] used the image of P. vivax parasite as their research object, in this case, to examine the result of the classification between segmentation on the saturation channel and the intensity channel from the HIS color space.…”

“…where n refers to the maximum value of grayscale and MSE refers to Mean Square Error, a value of the average error between the image before and after processing that can be counted using Equation (6). ( 6 ) where M, N refers to the resolution of the image, g(x,y) and g ' (x,y) refers to the value before and after processing. From the Equation (5), it can be seen that PSNR is inversely proportional with MSE.…”

“…by finding parasite visually in the red blood cells (erythrocytes). In such identification process, in addition to the good condition of the microscope, it must also be supported by the competence and experiences of the microscopists or the paramedics [6]. This method is still potential to lead to the misdiagnosis; therefore, given the today advance technology, such misdiagnosis can be gradually coped with.…”

Computational microscopic imaging for malaria parasite detection: a systematic review

“…On the other hand, thin blood smear with Leishman stain is considered in few studies (Table 4). It is reported that Leishman stain has good sensitivity for parasite detection than Giemsa (Khan et al, 2014). Giemsa stain is much costly and also time-taking procedure than Leishman.…”