Identify malaria parasites in blood smears using computer vision
Malaria is an acute disease caused by parasites of the genus Plasmodium. Commonly these parasites are transmitted to people through the bite of infected females. Among the first and most common symptoms are fever, headache and chills. They are generally difficult to recognize because they can be mild, appearing 10 to 15 days after the bite of the infecting mosquito. However, when the symptoms are not treated, they lead to a serious clinical picture that can lead to death in 24 hours (1).
One of the first tools implemented by WHO over the last two decades is the extension of access to malaria prevention strategies. Particularly effective vector control procedures and the use of preventive antimalarial drugs. This strategy has hardly helped to reduce the global burden of disease. Another basic malaria elimination strategy is vector control. This is very effective in preventing infection and reducing the transmission of the disease. The main explosions in this case are the use of insecticide-treated nets and indoor spraying with residual-action insecticides.
Chemoprophylactic treatments are those that use drugs in order to prevent malaria disease and of course its consequences. These include intermittent preventive treatment of lactating children and pregnant women, seasonal antimalarial chemoprophylaxis, and mass drug administration. This also includes the rapid diagnosis strategy in case of suspected infection and the treatment of confirmed cases. Finally, we find the vaccine since October 2021, which is administered to children living in areas with moderate to intense transmission. It has been proven that the vaccine significantly reduces the incidence of disease and mortality (1).
AI is changing the way healthcare services are delivered in various settings. Their processes have been facilitated by the increasing availability of large data sets and novel analytical methods that are based on these sets. Despite the multiple attempts to create prevention, diagnosis, and timely treatment strategies by the different health entities, there are problems such as the appearance of resistance to antimalarial drugs or the early detection of parasites that reduce the effectiveness of these strategies. This first creates the need for new alternative drugs, for which traditional drug identification approaches are time consuming and resource intensive (2).
Second, a light microscopic examination of blood smears is the gold standard technique for the diagnosis of malaria. However, this is a method that requires a lot of time and requires highly qualified personnel to perform the microbiological diagnosis (3). In recent studies AI has shown accurate performance using structure-based approaches in the field of chemical property prediction. Since some data already exists, AI would be a suitable tool for the identification of new drugs. These models might be able to learn patterns within the data and help identify successful and effective compounds. There are also new techniques based on the analysis of digital images using deep learning and artificial intelligence methods, which are defined as novel alternative tools for the challenging diagnosis of this infectious disease (3).