YOGA POSE VALIDATION USING DEEP LEARNING

YOGA POSE VALIDATION USING DEEP LEARNING

MS ROSHINI S.N(AP/AI&DS)

MR ARASU PANDIAN N, MS KRITHIKA R.S , MR KISHORE S

BACHELOR OF TECHNOLOGY -2ND YEAR

DEPARTMENT OF ARTIFICIAL INTELLIGENCE AND DATA SCIENCE

SRI SHAKTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY (AUTONOMOUS)

COIMBATORE - 641062

ABSTRACT:

The yoga practice was developed 5000 years ago in ancient India by the Indus- Sarasvati culture. The term yoga describes a deep union of the body and the psyche. It is used to maintain body and mental balance through all of life's ups and downs through asana, meditation, and other disciplines. Yoga has recently gained worldwide interest in the modern way of life due to increased stress levels, and there are numerous resources and methods available for studying it. Yoga can be performed in studios, with private instructors, and independently with the use of books, videos, the Internet, and other resources. Many people prefer self-learning because of their hectic schedules and the likelihood that the aforementioned materials won't always be available. However, with self-learning, a bad posture might not occur. Poor posture can have a negative impact on one's health, causing both short-term, acute pain and long-term, chronic issues. This research develops deep learning-based methods to identify bad yoga posture. Users can choose their preferred yoga pose by using this strategy. In order to train models that produce the anomalous angles found between the user's pose and the actual position, the user's pose is given. By highlighting the areas where the yoga stance is incorrect, the system uses these outputs to give the user advice on how to correct it.