What is SwiMe
SwiMe is a technology is based on a single camera with AI and computer vision. The software can detect the swimmers (angles, joint trajectories, stroke rate, and more). We are in beta mode and can only analyze Freestyle/Front Crawl taken using an underwater camera in two views: front and side. Users can upload movies free of charge and get feedback in approximately 1-24 hours. Currently, the app is designed for a computer screen
How it Works
Record a Video
- Use waterproof camera
or phone - Place it approximately 40cm, under the water surface.
Record the swimmer swimming to the camera, or side view - Videos should be smaller than 100Mb (typically 5-10sec) in mp4 or mov format
Upload
- After recording your movie, check that it is similar to the example movie below
- You will upload a movie
- If you are a swimmer you will load your own
- If you are a coach you can upload your swimmers’ movies
View
- Our AI will analyze your strokes
- All the movies you have uploaded will be stored.
- You can also share your videos
- Currently, the app is designed for a computer screen
- Use Google Chrome
- Learn More – Github Page
Let SwiMe Do It's Magic
About SwiMe
SwiMe is a research project from Ben Gurion University. Motion is a key factor in activities such as sports and rehabilitation that depend on performing the proper motion in the right way (e.g., physiotherapy and ergonomics). Thus, many coaches and researchers analyze their swimmer videos. However, this is a manual process that is very time-consuming and subjective. Our software can calculate main swimming parameters such as stroke rate, kick rate, and present the joint location and angle trajectory as a function of time, and more. All these parameters, including the video, could be downloaded for your own use. Since we are in development/research mode currently, the system can process freestyle very well, and butterfly (with less accurate results). You are welcome to test and use this system
The People Behind SwiMe
Raziel Riemer Ph.d
Biomechanics and wearable robotics.
Certified swimming coach
Former Israeli national swimming champion.
Gera Weiss Ph.d
Mathematics and Computer Science.
An amateur master swimmer
Itay Coifman Gaon
R&D
Data-driven biomechanics
Machine Learning, Data Science
May Hakim
R&D
Human motion analysis.
Machine Learning
Contact Us
FAQs
Additional informatin about the possible outputs is found in the dedicated Github Page
Type should be mp4
The system can deal with up to 100Mb
Start the video when the swimmer is approximately in the middle of the pool (in 25m/y).
This should be good for 5-15 sec movies
A coach can upload movies for his swimmers and save them by the swimmer name
you can then send a link to your swimmers with the video
Swimmers can share their videos with their friends
Currently we are testing mode and each movie is evaluated manually by an expert before release.
Thus, it will take between 1 to 2 days.
In the future this will be less than a minute.
It is best to use Google Chrome, other Internet Browsers might not work