The platform utilizes a highly sophisticated stack including agentic AI, computer vision, deep learning, and machine learning.
An innovative Artificial Intelligence model capable of acting as a personal fitness instructor has been developed to identify and correct human postures during yoga sessions. Dubbed the ‘Yoga AI Coach,’ the system is slated for a massive scale trial during the upcoming International Yoga Day on June 21, where drones will capture live footage to test its tracking accuracy on large crowds.
Developed under a research initiative at the Shri Govindram Seksaria Institute of Technology and Science (SGSITS) in Indore, the project has secured a grant of 7.10 lakh rupees from the Madhya Pradesh Council of Science and Technology (MPCOST). The system is designed to provide immediate feedback, making correct yogic practices accessible to global audiences who lack access to professional trainers.
Technologically, the platform utilizes a highly sophisticated stack including agentic AI, computer vision, deep learning, and machine learning. By analyzing body angles with extreme precision, the model can instantly flag even a minor deviation in arm or leg positioning.
Beyond basic posture correction, the software offers advanced analytics, including real-time facial expression mapping, performance scoring, and gender-based participant counts. The model has been trained on a massive dataset of approximately 2.80 lakh images, allowing it to evaluate 104 distinct yoga poses, including the complex Sun Salutation (Surya Namaskar).
The development team plans to offer the application entirely free to the public once the necessary hardware infrastructure and long-term financial maintenance support are secured.
