Our History
2016 – First prototype of a motion tracking fitness application was completed using HAAR Cascade classifiers, limited to push-ups and simple movements such as squats and shoulder press exercises. Ability to count repetitions developed.
2018 – First development of deep learning models trained using videos of the developer’s workouts using a machine artificial intelligent compute engine. Q3, First release of a mobile motion tracking fitness application on Google Play under Science Fitness.
2019 – Retrained AI models were created with a more accurate training methods for faster performance and greater compatibility with mobile devices.
2020 – Extension of AI models began training from just simple barbell and dumbbell exercises, into body exercises. AI model trained to determine depth with 3D pose estimation from a single source camera.
2022 – AI models further trained additional sports and health activities such as Yoga, Aerobics, Boxing, Skipping and much more. Motion tracking of body developed the ability to determine speed, posture and AI ability to correct posture with real-time response.