Experiments with Real-time Human Pose Estimation

With the emergence of artificial intelligence and the evolvement of computer hardware, computer vision algorithms are getting better and better. The work is using a Real-time Human Pose Estimation algorithm (PoseNet) that can detect human figures in images and video. To be clear, the algorithm does not know ‘who’ is in the image, but it is estimating where key body joints are in the given image. The videos show scenes of well known movies in an unusual mannor. Algorithms that make it able to abstract what we see to make up the human body. Questioning, and reflecting on, what our brain needs to make sense of what we see.