Our Malaria Detection Application
Our purpose is to disrupt the current process for Malaria diagnosis that is labor-intensive. We want to provide quick and accurate detection of Malaria parasites in patients to regions that lack the access to microsopists.
Our Automated Malaria Early Detection System application is a mobile application that can be downloaded onto a smart phone and can process the image of any blood slide under a 100x objective lense microscope to determine what Malaria treatment should be administered for the patient.
Motivation
Average Annual Deaths from Malaria
Malaria is curable if the patient is treated in time.
False Diagnosis of Malaria in Blood Smear Test
Average time spent diagnosing a slide by a microscopist
With these dire statistics, our team is motivated to make an impact to countries(large portions of Africa and Asia) and vulnerable victims (young children and pregnant women) hugely impacted by the Malaria crisis with our application where Malaria diagnosis can have less false diagnosis with its accuracy. With this application, we intend to reduce the time overhead with its speed to deliver Malaria test results quicker to patients and therefore ensures the patient is treated in time.
Design and Deliverables
To achieve our application, we used the following technologies:
Machine Learning
From training a machine learning model, we can diagonse a entire blood slide for Malaria in under 5 seconds. With our test data set from the National Institutes of Health (NIH), we have an accuracy of 97%
End-to-end Classification
The application can identify and seperate individual blood cells from a single blood slide. After abstraction, it can classify which blood cells are infected by the Malaria parasites.
Mobile Application
To achieve portability, our application can be used on any modern-day smartphone. Download the application, simply attach the phone's camera to the eyepiece of a microscope with 100x objective lense.
Cloud Hosting and Service
To increase both the portability and accuracy of our application, we have worked behind-the-scenes to create a data pipeline to remove the storage space on your phone and provide you with updated machine learning models.
Awards
Here is the growing list of competitions and grants that we have won at this moment in time:
Our Team
Our team is composed of undergraduate students with representation from UC Berkeley and San Francisco State University interested in designing products that have real-world impact.
Wanzin Yazar
Machine Learning Lead
Paulina Nguyen
Product Manager
Austin Tsang
Full Stack Lead
Phyo Khine
Mobile Developer