An interactive mindmap of blood film images for automated malaria diagnosis: a comprehensive metadata repository of available datasets

Charles B. Delahunt, Itunuoluwa Isewon

Abstract: Summary: Malaria remains a significant public health challenge, necessitating innovative diagnostic solutions to combat high mortality rates. Here, we conducted a systematic review of malaria microscopy datasets published between 2014 and 2024 by querying Scopus and PubMed databases. We have a curated inventory of 71 malaria blood film image datasets. To further enhance usability, we developed an interactive mind map that visually represents the datasets based on five key features. The Access category differentiates between open and controlled access datasets. The Smear Type category classifies datasets by blood smear type. The Stain Type category organizes datasets by the staining methods used. The Species Type highlights the Plasmodium species present in the images, while the Demography category identifies the geographic origin of datasets. Each dataset has a dedicated page with a detailed description, a citation of the dataset's publication, and links to databases where it can be downloaded. This mind map has been carefully curated with up-to-date information and will be regularly updated as new datasets become available, ensuring its continued relevance and accuracy. The mind map is a freely accessible resource, available at https://itunuisewon.github.io/Malaria_Blood_Film_Images/, which enables researchers to identify and utilize diverse datasets for training and validating machine learning algorithms.