2D Posture Estimation of Human Arm

• Published in: International Conference on Mathematics and Computers in Sciences and in Industry (Online) pp. 168 - 173
• Main Authors: Rosca, Sebastian Daniel, Leba, Monica, Sibisanu, Remus
• Format: Conference Proceeding
• Language: English
• Published: IEEE 22.08.2024
• Subjects: Arms; Biological system modeling; Biomimetics; control; Kinematics; Machine learning algorithms; Manipulators; Mathematical models; Pose estimation; Regression tree analysis; robot; Service robots
• DOI: 10.1109/MCSI63438.2024.00035

Human pose estimation and automatic detection is a common task in Human-Computer Interaction but especially more capitalized in human-robot interaction inspired from biological form of human anatomy to realize a biomimetic control. In this sense, using human pose estimation it can be realized the whole human body posture estimation using anatomical human joints detection as inputs for specialized algorithms to calculate it based on key points obtained from a person Kinematic tree using regression, generation or part relation as models to estimate the optimal position from video or even from image. In this paper we propose a solution to estimate of 2D human arm posture of a subject using image capture as input for advanced machine learning algorithms to determine the subject joints position used to control in biomimetic manner a simulation of human like robotic arm with six degrees of freedom as a computer-aided model in Matlab-Simulink based on forward kinematics.