Proposal of a Metaheuristic Algorithm of Cognitive Computing for Classification of Erythrocytes and Leukocytes in Healthcare Informatics
Ana Carolina Borges Monteiro, Reinaldo Padilha França, Rangel Arthur, Yuzo Iano · 2022
Through the use of AI in diagnostic medicine provides accurate and safe diagnoses in the evaluation of exams. Image recognition is not an easy task, as this can be achieved with the ability to organize images in an automated way provided by machine learning, classifying this based on identified patterns, particular objects, and grouping them thematically. In this panorama, pattern recognition distinguishes issues effectively, through a voluminous database containing data related to the pathology, enabling machine “learns” to do an accurate diagnosis of the patient's condition. In this context, through blood count through manual and/or automated procedure, blood is investigated, motivated by innovations in the field of medicine, implementing a Deep Learning framework for the recognition and identification of white blood cell subtypes in digital images, it is possible to employ employing a set of convolution layers, allowing to distinguish details not revealed to the human naked eye, extracting resources (from edges) from WBC digital images molding the convolution operation, also relating feed-forward layers combined for WBC augmented digital image for cell classification and qualification. With this focus, using Python language and Jupyter notebook software, the dataset encompassing 12.500 digital images of human blood smear was manipulated, integrating fields of non-pathological leukocytes. Resulting from that, with an accuracy of 86.16% testifying the elevated reliability of the developed framework. Thus, the proposed framework is evaluated as an accurate, low-cost, and effective digital method that can be used as a third practicable procedure for blood count in the underprivileged populace of underdeveloped and developing countries.