Inspiration
Ethan Zhu's mother is a physician who has mentioned that certain clinical diagnosis that deals with body deformities take a long time to diagnose due to manual body measurements. So our take on utilizing computer vision to automate the process of measuring the ratio between one's limbs and their height.
What it does
With 4 individual pictures, the python program can identify an individual's face, eyes, feet, shoulders, and hands. Following this, the program will then determine the ratio between the length of a person's arms and their height, which can be indicative of a body deformity, requiring a medical diagnosis.
How we built it
The python program was built using open source computer vision algorithms in OpenCV. Haar feature-based cascade classifiers were used to identify facial features, while template matching was used to determine the location of the feet, hands, and shoulders.
Challenges we ran into
Completing our project on time required intensity and perseverance, which was difficult because of our lack of sleep. However, we overcame this by keeping our group's energy up, and ensuring that we all felt motivated. Additionally, reducing the dimensionality of the algorithm so that it produced one selection per feature was challenging.
Accomplishments that we're proud of
Implementing our hack's feature selection with a high degree of accuracy, using a platform new to all members of our team.
What we learned
OpenCV, computer vision software, algorithms, and how data is stored and produced by packages.
What's next for MediMeasure
A mobile application accessible to any physician which can easily and simply take the desired photos and analyze them for any deformities.
Log in or sign up for Devpost to join the conversation.