Automatic analysis and diagnosis of disease
This computer science project aims to provide medical institutions with an alternative method to diagnosing diseases that require a blood sample to be tested and analysed.
We will use peripheral blood smear images and monitor the affects of a disease on red blood cells and their morphology. Changes in cell size, shape, colors, et cetera can relate to a specific disease so we might be able to make a diagnosis based on the information contained within the sample image. For an extensive explanation of the problem description, related work, methodology, user requirements and more please the supplied documentation.
Diseases such as Tuberculosis, HIV/AIDS, and Diabetes kill 500 000 people every year in South Africa and affect millions of people world wide. Diagnosing these kinds of diseases is a two-part process. Tests need to be conducted on a sample belonging to the patient, and the test results need to then be anaylsed by a specialist medical practitioner. We aim to eleminate the gap between the testing and analysis/diagnosis procedure by creating a software product. Our solution will attempt to automatically analyse blood samples and then provide a diagnosis of disease in a timely, less costly, more easily and efficient way.
TOP KILLING DISEASES IN SOUTH AFRICA:
Influenza and Pneumonia
Cerebrovascular and other Heart Disease
Student: Mathew Ramsay
Dr. Jean-Baka Domelevo Entfellner, Mehrdad Ghazi-Asgar, Reggs Dodds, Dr S. Tondeur
For more information contact the Computer Science department at the University of the Western Cape