Success case
Computer Vision with Fitness: Effectively Evaluate Human Motion with Feedback in Real-time
Hardware: NVIDIA Jetson Orin Nano Application: Pose Estimation for Gym Training Industry: Intelligent Fitness Evaluation Deployment Location: US, Spain, France In today’s fast-paced world, fitness enthusiasts and athletes are constantly seeking innovative ways to enhance their training routines. It’s getting more critical to learn precise and real-time feedback during gym workouts. Traditional methods of assessing […]
- Industry
- Retail & Consumer Services
- Application
- Gym training pose evaluation
- Results
- 106 FPS Inference speed
Project background and needs
This pose evaluation solution is deployed in gyms across the US, Spain, and France, using Jetson Orin Nano and IP cameras to analyze training movements and provide feedback.
Subjective observation
- Traditional pose evaluation relies on the naked eye.
- Evaluation is prone to subjective judgment and human error.
Needs consistent pose analysis results.
Feedback timeliness
- A coach may supervise multiple clients at the same time.
- It's hard to flag each person's form deviations promptly during a workout.
Needs real-time training feedback generated from video.
Solution and architecture
Vision inference
IP camera images are processed by the YOLOv8-s pose model and TensorRT, trained on the COCO-pose dataset.
Motion feedback
Body motion data identifies the training type and counts repetitions, giving coaches a basis for guidance.
Solution components
On-site gallery (2)
Discuss your application with our team
Tell us about your requirements and deployment environment.