One of the features that makes the BitGym cardio app special is the ability to exercise through beautiful scenery and feel truly connected to the place, by having the scenery move at the pace of your cardio exercise. Research has shown that immersion is increased through interactivity (sensory feedback closely tied to your physical activity) and increased immersion leads to increased enjoyment, motivation, and mental and physical benefits.
This is possible because of BitGym's patented technology for detecting your exercise speed without any proprietary connection to your exercise machine, or wearing any extra hardware sensors. We achieve this by using a sensor available to anyone with a modern smart-device: the front-facing camera, and the processing power on your phone/tablet itself. In this article you will find a simple overview of how we do this and the answers to some common questions.
1. What sensors do you use?
We use only the front-facing camera to detect your exercise. In previous versions, we detect if your device is shaking too much using the accelerometer/gyrometer to warn you the tracking may be unstable, but now we accomplish this with your camera as well.
2. Do you record anything?
Nope. No video data (image pixels, colors, timing, audio, compressed video, encoded or decoded) from your device camera is saved to your device or to our servers during BitGym's normal operation.
3. Do you process and store any information from the camera?
The only information we process is your exercise cadence, which is stored right in your workout record and graph for you to see. Essentially, the only information from your camera we store is what you can see. When you delete a workout record, it's deleted completely from your device and from our servers.
4. Don't you need to send camera information to your server to perform the tracking?
We don't! BitGym tracking happens entirely on your mobile device's processor, in real-time, completely offline. This is why you can use BitGym's tracking even without any internet.
5. How do you improve your tracking technology then? Don't these techniques collect lots of data to improve the tracker?
Unlike computer vision technology described in modern popular news (self-driving cars, image classifiers, etc), BitGym's method does not rely on large datasets of collected images or video and large server farms processing these to train data models. Rather, it is based on relatively simple signal processing on "optical flow" motion. This is why it is possible on mobile devices, with minimal processing time, and no datacenters full of user video.
When we do see the need to improve the tracking technology, we request users to fully voluntarily