• Dr Rashmi P.Sonar
• Sakshi Taywade
• Devashri Khorgade
• Nikita Naik,
• Anjali Bonte

: This article describes a novel approach to the classification of whole-body motions from estimated human postures in 2D camera images and the subsequent generation of their relevant description. Action description from 2D human body postures is a challenging task with a wide range of applications, such as yoga instruction, yoga rehabilitation, and yoga research. Previous approaches to this task have relied on hand-crafted features or shallow machine learning models, which have limited accuracy and scalability. This paper proposes a new approach to action description using deep-learning image classification. The proposed approach uses a deep learning model to extract features from 2D human body posture images and then classify the images into different actions. The model is deployed on a Django web application, which allows users to upload 2D human body posture images and receive a description of the action they are performing. The proposed approach is evaluated on a large dataset of yoga poses.

Action description, yoga poses, yoga information, deep learning, image classification, Django web application.

"Action Description From 2D Human Body Postures ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 10, page no.h652-h657, October-2023, Available :http://www.jetir.org/papers/JETIR2310681.pdf

"Action Description From 2D Human Body Postures ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 10, page no. pph652-h657, October-2023, Available at : http://www.jetir.org/papers/JETIR2310681.pdf