Poor posture in the workplace can lead to chronic pain, decreased productivity, and long-term health issues. Implementing reminders for proper posture is crucial to enhance employee well-being, reduce discomfort, and improve overall efficiency. The app is designed to detect poor posture using a pose detecting neural network and provide immediate feedback to the user. When it identifies slouching or improper alignment, it sounds an alarm and sends a notification through a sleek, user-friendly interface. This real-time alert system encourages users to correct their posture instantly, helping to prevent discomfort and long-term health issues. As well, proper posture is rewarded with a green background, giving positive reinforcement to improve posture techniques.
We used the HRNetPose model from Qualcomm AI Hub to run our app entirely on-device with Snapdragon X Elite!
Developed by:
- Thomas Lascaud, tlascaud@qti.qualcomm.com
- Ashley Nicole Leal, aleal@qti.qualcomm.com
- Lucie Yang, luciyang@qti.qualcomm.com
- Laura Lin, laurlin@qti.qualcomm.com
- Keshavadithya Subramanya, keshavad@qti.qualcomm.com
- Python >=3.9 and <=3.11 installed
- Node.js installed
- npm installed
- git installed
git clone https://github.com/ashleyleal/haQathon.git
cd haQathon
Download hrnet_pose-hrnetpose-float.onnx from here: https://aihub.qualcomm.com/compute/models/hrnet_pose and put it in pose-detect
Choose ONNX Runtime > float
Install dependencies
cd pose-detect
python -m venv venv
source venv/bin/activate # on Windows use: venv\Scripts\activate
pip install -r requirements.txt
Run the backend server
fastapi dev backend.py
Run
cd frontend
npm install
npm run dev
Go to http://localhost:3000
Derrick Johnson's example: https://github.com/DerrickJ1612/qnn_sample_apps/blob/main/notebooks/pose_detection/Pose-Detection-with-HRPoseNet.ipynb