Fuzzy Posture Matching for Pain Recovery Using Yoga

Owing to the present sedentary lifestyle, most of the office employees suffer from pain at different body parts and in turn they need to consult with doctors frequently. The doctors prescribe certain exercises (mostly in the form of yoga) to the patients to overcome the stiffness and pain. Yoga is an ancient practice which involves specific movements, and breathing exercises used to improve overall health. With the rapid advancement in computer vision and video processing technologies, automatic detection of yoga postures’ correctness will help to motivate the patients to do them accurately. Here lies the novelty of our proposed work of automatic detection of yoga postures for pain recovery using type-1 fuzzy set. As care must be taken that the exercises are done appropriately, so a doctor-consulted database of yoga postures has been taken into account. The postures are detected by Kinect sensor first; then the recognized skeletal forms of the postures are processed to obtain the feature space. While performing the yoga, the exercise pattern may vary for different subjects due to their body structure. So to deal with this uncertainty, fuzzy logic has been used here. The proposed system is an innovative work in this domain and outperforms all the existing ones in other posture recognition domains also.