Aveo 2.0
02 December 2020
Aveo 2.0: Color Matching-based Macroscopic Urinalysis
Originally, Aveo is the codename for the project. My Team and I, participated in Imagine Cup 2009 and have created a project named Aveo. Aveo is a very big solution to address maternal mortality in the suburb area. We went to semifinal to present the works. The project was published to Microsoft Indonesia’s blog, but since the domain has expired, the project detail was gone as well.
And then, I revitalized the project into smaller project for my undergrad thesis. I picked the urinalysis system as topic. Previously, this part is done manually by user. Originally this part of solution is an embedded system solution that will be deployed in rural area. User takes one strip of urine test strip , and dip it into the urine. For several minutes, the color of strip will react accordingly. User will match the color with the table, and can judge whether has maternity problem or not.
I don’t want it to be manual. I want it to be automatic. So here it is, Aveo 2.0. FYI, this project was conducted at 2011. Some of technology might be obsolete.
Background Problem
Telemedicine is a mechanism to provide health service remotely. This can be applied anywhere, especially rural area. Some rural area may have unproper or limited access to health facilities. With technology, providing telemedicine to those rural people is not impossible. As started, I targeted pregnant mom as the user for my research. As we know, pregnant woman must put extra care to the baby, by eating good food, maintaining own’s health, etc. Research said that certain level of chemical in the body may harm the baby, or death for the worst.
Fortunately, an early detection can be done through urinalysis. I utilized computer vision technology to detect the color and match it automatically with the urine table. Once it is developed, the solution can be used anywhere and anytime. No need to visit hospital to check.
Proposed Solution
I utilized OpenCV 2 as vision library. I create an algorithm to detect the color of the strip and generate its value. After that I match the value to the reference urine table. This solution needs camera to detect the color strip. This solution is far from perfect. There are a lot of constraint to consider, such as light, and other environment that affect the color of the strip.
Thanks to this project for giving me my undergraduate degree~