Multi Person Pose Estimation

Pose estimation is a computer-based diagnostic technique that detects and evaluates a person's body shape. An important part of measuring a person's posture is the shape of the human body. Volume-based model, skeletal model and contour model are the three most commonly used types of human body models. Analysing human activities is beneficial in many areas such as video surveillance, biometrics, assisted living, home health monitoring etc. With our fast-paced lifestyle these days, people often prefer to exercise at home but feel the need for a teacher to evaluate their work. form. Since these resources are not always available, personal recognition can be used to create a self-study program that allows people to read and practice the tests effectively on their own. Our project forms the basis for building such a program by discussing various learning tools and in-depth learning methods to accurately distinguish yoga posture from pre-recorded and real-time videos. The project also discusses a variety of posture and methods for obtaining key points in detail and describes the in-depth learning models used for pose segmentation.