Yoga Teacher is a great app for practicing yoga at home. It uses a smart device camera and a sophisticated motion recognition algorithm to interactively monitor and measure the accuracy of asanas performed by the user.
The app will be an interactive training solution for yoga beginners, helping them improve their skills through practice.
The self-trainer app uses a device camera to track the user’s posture and movement, then compares it against multiple parameters of a pre-programmed asana and calculate the accuracy percentage.
We had to identify the most suitable method for posture tracking through the camera. The algorithm has to compare positions of matching points on the user’s body and the reference body, then calculate the accuracy percentage for the asana.
We had do discard several formulas due to issues with discerning a body from the background. For example, we used a Euclidean metric formula to calculate distances from corner points and determine the posture based on the results, but we had to reject this method later.
The technology in the product must now only discern the body from the background properly, but also do it in real time, without making the user wait.
Creation of a minimum viable product
Once we have developed the technology sufficient for an MVP, we tested it on a number of real yoga exercises, involving both beginners and advanced practitioners. In order to improve the quality of posture recognition, we made a decision to distinguish between male and female users.
Our team has conducted 3 iterative rounds of user testing, with interim evaluation and correction of parameters. We have managed to eliminate defects and achieve maximum accuracy in asana recognition.
Our specialists have studied an OpenCV computer vision library and found methods to implement it into iOS. The incorporation of recognition module into iOS was not an easy task. The technology was based on C++, which is not an OOP language, and this created conflicts with objective-c.
The app has to be easy to use, with simple and comfortable user interaction.
The visual design has to be pleasing to the eye, but at the same time inspire users to practice yoga.
Based on this, we prepared 3 concepts:
The app contains 24 fully functional exercise sets for beginners and advanced practitioners to attain their fitness goals, e.g. weight loss. After each exercise, asana accuracy graphs are displayed, making it easy for the user to track their progress through practice. A calendar is included for recording the practice history and results as well as for planning future practice sessions.
Requirements of our clients at the stage of project evaluation
- Successful experience in developing projects for medium and large businesses.
- Business expertise in the development of start-ups.
- A team of 7 specialists when developing an application or a website.
- Uniqueness, no standard solutions.
- Term - development from 3 months.
- Transparent control system and flexibility in development.
ZealStep is a unique application where virtual money is assessed for complete steps and which can be exchanged for goods and services, or spent on getting discounts in the gyms, fitness centers, on buying items for sport and health.
Project for the existing in Switzerland taxi business. The project includes a mobile application for passengers under iOS and Android, an application for drivers for iOS and an administrative panel.