Malaria is an endemic life-threating disease caused by the unicellular protozoan parasites of the genus Plasmodium. Confirming the presence of parasites early in all malaria cases ensures species-specific antimalarial treatment, reducing the mortality rate, and points to other illnesses in negative cases. However, the gold standard remains the light microscopy of May-Grünwald–Giemsa (MGG)-stained thin and thick peripheral blood (PB) films. This is a time-consuming procedure, dependent on a pathologist’s skills, meaning that healthcare providers may encounter difficulty in diagnosing malaria in places where it is not endemic. This work presents a novel three-stage pipeline to (1) segment erythrocytes, (2) crop and mask them, and (3) classify them into malaria infected or not. The first and third steps involved the design, training, validation and testing of a Segmentation Neural Network and a Convolutional Neural Network from scratch using a Graphic Processing Unit. Segmentation achieved a global accuracy of 93.72% over the test set and the specificity for malaria detection in red blood cells (RBCs) was 87.04%. This work shows the potential that deep learning has in the digital pathology field and opens the way for future improvements, as well as for broadening the use of the created networks.
9 May 2018 in International Journal of Laboratory Hematology
Objectives: The objective was the development of a method for the automatic recognition of different types of atypical lymphoid cells. Methods: In the method development, a training set (TS) of 1,500 lymphoid cell images from peripheral blood was used. To segment the images, we used clustering of color components and watershed transformation. In total, 113 features were extracted for lymphocyte recognition by linear discriminant analysis (LDA) with a 10-fold cross-validation over the TS. Then, a new validation set (VS) of 150 images was used, performing two steps: (1) tuning the LDA classifier using the TS and (2) classifying the VS in the different lymphoid cell types. Results: The segmentation algorithm was very effective in separating the cytoplasm, nucleus, and peripheral zone around the cell. From them, descriptive features were extracted and used to recognize the different lymphoid cells. The accuracy for the classification in the TS was 98.07%. The precision, sensitivity, and specificity values were above 99.7%, 97.5%, and 98.6%, respectively. The accuracy of the classification in the VS was 85.33%. Conclusions: The method reaches a high precision in the recognition of five different types of lymphoid cells and could allow for the design of a diagnosis support tool in the future.