Aegle

HeadSittingScore3 - Head Upright
HeadSittingScore0 - Head tilted
TrunkSittignScore1 - Spine slightly curved
Assessing real time camera feed from user

Inspiration

Cerebral Palsy (CP) affects about 1 in 345 children and is the most common motor disability in childhood and considered to be the leading cause of childhood disability worldwide. Early diagnosis has significant benefits including minimization of complications such as hip dislocation, scoliosis and spasticity. Despite this, assessment for CP can take 18-24 months missing the crucial rapid neural development stage for intervention. Cerebral Palsy assessments are also not be as widely accessible as they are lengthy and must be done over long periods of time. Aegle aims to change that.

What it does

Aegle leverages machine learning in conjunction with the HINE evaluation model to determine early markers of Cerebral Palsy. The app captures live video to analyze markers of Cerebral Palsy. These markers are then compiled to produce a HINE score, in which Aegle makes it easy to export the score to be further examined by a licensed physician. and get the needed support. Assessments are recommended to be conducted over the course of 9, 18, 30 and 48 month marks of the child's age.

How we built it

I used the Mediapipe library to train my model and determine landmarks for poses, facial expressions and hand gestures that follow the HINE evaluation model. The code is written on python and runs on Google Colab.

Challenges we ran into

During the building process I had problems training my model and ensuring that it was able to identify certain poses. In addition to this, I often ran into bug issues with the code.

Accomplishments that we're proud of

I'm proud that I was able to do create a project that is meaningful and impactful. I'm also proud that I was able to successfully train my models and produce results.

What we learned

I learned more about the Mediapipe library as well as further python skills.

What's next for Aegle

Aegle hopes to continue this project and train their models with children so that they are accurate and valid. We also hope to add additonal assessments to our project to compliment the existing HINE evaluation model and provide a more accurate, fullsome score for our users.