Real-Time Human Pose Estimation: A MediaPipe and Python Approach for 3D Detection and Classification

As the name suggests, this study uses the Open- Pose and MediaPipe frameworks in order to give a thorough analysis of real time human posture detection and identification. Predicting the locations of key body joints in a person's stance is a typical computer vision issue known as human posture identification." We use the strength of OpenPose and MediaPipe to quickly and reliably identify different positions. Our approach concentrates on particular stances, such as the T-pose, Tree-pose, and Warrior pose, and gives users visual feedback while watching a live camera feed. The system analyses and recognizes postures using a structured output of body key points, confidence ratings, and pixel coordinates. The project runs well under ideal lighting settings and with adequate hardware resources, while it may occasionally lag and show inconsistent behavior under difficult circumstances. Through algorithm optimization and hardware upgrades, future work can overcome these restrictions and improve the system's performance. With potential applications in fitness tracking, virtual reality, and rehabilitation, this study advances the field of human posture detection and recognition.