Skeletonization techniques for object detection - state-of-the-art report
This work presents the results of the literature research on the subject of recently published approaches concerning skeletonization algorithms for object detection. It gives a short survey of the principles of skeletonization and object detection. The main part is focused on the state-of-the-art of generating a skeleton of a human with additionally poses estimation from an input depth image. At the end a brief conclusion including a discussion of advantages and disadvantages of the several presented approaches is given.