Mobile-Based Analysis of Malaria-Infected Thin Blood Smears: Automated Species and Life Cycle Stage Determination

Abstract

Microscopy examination has been the pillar of malaria diagnosis, being the recommended procedure when its quality can be maintained. However, the need for trained personnel and adequate equipment limits its availability and accessibility in malaria-endemic areas. Rapid, accurate, accessible diagnostic tools are increasingly required, as malaria control programs extend parasite-based diagnosis and the prevalence decreases. This paper presents an image processing and analysis methodology using supervised classification to assess the presence of malaria parasites and determine the species and life cycle stage in Giemsa-stained thin blood smears. The main differentiation factor is the usage of microscopic images exclusively acquired with low cost and accessible tools such as smartphones, a dataset of 566 images manually annotated by an experienced parasilogist being used. Eight different species-stage combinations were considered in this work, with an automatic detection performance ranging from 73.9% to 96.2% in terms of sensitivity and from 92.6% to 99.3% in terms of specificity. These promising results attest to the potential of using this approach as a valid alternative to conventional microscopy examination, with comparable detection performances and acceptable computational times.

Keywords: computer-aided diagnosis; image analysis; malaria; microscopy; mobile devices.

Figure 1

Blood smear analysis flow for both quantification and species/life cycle stage identification.

Figure 2

Mobile-based framework for malaria parasites’s detection: (A) $\mu$SmartScope with smartphone attached and blood smear inserted; (B) smartphone application screenshots; (C) exemplificative usage of the solution (from left to right): (i) blood smear insertion; (ii) start image acquisition through the smartphone app; and (iii) visual feedback of the automated detection.

Figure 3

Diagram of the proposed methodology for the automatic analysis of thin smear images.

Figure 4

Illustrative examples of different MPs species and life cycle stages from the Mobile Thin Smear Malaria Parasites (mThinMPs) database.

Figure 5

Effect of brightness and contrast adjustment, with cumulative histograms: (A) original image; (B) processed image after $\alpha$ and $\beta$ correction, followed by mean-shift filtering.

Figure 6

Pre-processing: (A) original image; (B) brightness and contrast adjustment; (C) sharpening applied over green channel of adjusted image; (D) RBCs; segmentation applied over the sharpened image; (E) blue channel of the original image; (F) optical circle segmentation applied over the blue channel of the original image.

Figure 7

Examples of trophozoites ring stage candidates: (A) original image (cropped ROI); (B) brightness and contrast enhancement; (C) cytoplasm grayscale sharpening; (D) cytoplasm segmentation and filtering; (E) chromatin grayscale sharpening; (F) chromatin segmentation and filtering; (G) final candidates (cytoplasm in red; chromatin in yellow; RBC with candidate inside in green).

Figure 8

Examples of mature trophozoite stage candidates: (A) original image (cropped ROI); (B) brightness and contrast enhancement; (C) cytoplasm grayscale sharpening; (D) cytoplasm segmentation and filtering; (E) chromatin grayscale sharpening; (F) chromatin segmentation and filtering; (G) final candidates (cytoplasm in red; chromatin in yellow; RBC with candidate inside in green).

Figure 9

Examples of schizonts candidates: (A) original image (cropped ROI); (B) brightness and contrast enhancement; (C) cytoplasm grayscale sharpening; (D) cytoplasm segmentation and filtering; (E) merozoites’ chromatin grayscale sharpening; (F) Merozoites’ chromatin segmentation and filtering; (G) final schizonts candidates (cytoplasm in green; chromatin in yellow).

Figure 10

Gametocytes candidates: (A) original image (cropped ROI); (B) brightness and contrast enhancement; (C) grayscale sharpening; (D) segmentation and filtering; (E) final candidates (at green).

Figure 11

Illustrative examples of the data augmentation procedure.

Figure 12

Examples of false negatives’ candidates for different species and life stages after segmentation and filtering.

Figure 13

Heat maps of the SVM parameters’ selection process for each species-stage combination.

Figure 14

Classification models workflow. (a) Diagram of the classifier models workflow for the detection of multiple species-stage combinations in a single image. (b) Illustrative examples with detection of: (I) P. falciparum trophozoites and gametocyte; (II) P. ovale trophozoite and gametocyte; (III) P. malariae trophozoites and schizonts.