Many employees who work at desks in this increasingly digitized world have bad neck and back posture. This bad posture has contributed to the prevalence of spine-related health issues in our society. To fix this, ergonomic issues must first be identified. A computer-vision algorithm can easily examine the structure of a person sitting in a quick and efficient manner, evaluating people's working poses to better workplace health.
What it does
ErgoPose takes in a right or left side profile of a person in a normal working sitting position and examines key angles of the spine, neck, and legs. After analyzing angles and comparing to an ideal set of working positions, ErgoPose provides easy-to-access and easy-to-implement suggestions for bettering posture.
How I built it
OpenPose API, Python, and Google Cloud
Challenges I ran into
Finding a suitable API, coding ideal positions
Accomplishments that I'm proud of
Analyzing computer vision results, producing personalized suggestions, building a functional model within Google Cloud
What I learned
Using Google Cloud to store and leverage data sets, understanding APIs, analyzing data and apply it to ergonomic concepts
What's next for ErgoPose
We would like for companies to use our hack to provide personalized ergonomic suggestions rather than general advice. We are also looking into applying this to other aspects of our daily lives, such as exercise or sleep.