Real-Time Pose Estimation from Images for Multiple Humanoid Robots

Pose estimation commonly refers to computer vision methods that recognize people’s body postures in images or videos. With recent advancements deep learning, we now have compelling models tackle the problem real-time. Since these are usually designed for human images, one needs adapt existing work on other creatures, including robots. This paper examines different state-of-the-art pose and proposes a lightweight model can real-time humanoid robots RoboCup Humanoid League environment. Additionally, present novel dataset called HumanoidRobotPose dataset. The results of this potential enable many advanced behaviors soccer-playing