Estimating Kinematic Structure

Unsupervised Learning of Complex Articulated Kinematic Structures combining Motion and Skeleton Information

Hyung Jin Chang and Yiannis Demiris
Imperial College London



In this paper we present a novel framework for unsupervised kinematic structure learning of complex articulated objects from a single-view image sequence. In contrast to prior motion information based methods, which estimate relatively simple articulations, our method can generate arbitrarily complex kinematic structures with skeletal topology by a successive iterative merge process. The iterative merge process is guided by a skeleton distance function which is generated from a novel object boundary generation method from sparse points. Our main contributions can be summarised as follows: (i) Unsupervised complex articulated kinematic structure learning by combining motion and skeleton information. (ii) Iterative fine-to-coarse merging strategy for adaptive motion segmentation and structure smoothing. (iii) Skeleton estimation from sparse feature points. (iv) A new highly articulated object dataset containing multi-stage complexity with ground truth. Our experiments show that the proposed method out-performs state-of-the-art methods both quantitatively and qualitatively.


Paper | Extended Abstract | Poster


[Dataset & Code]

Imperial-PRL-Dataset [download]
Source Code [download]


author = {Hyung Jin Chang and Yiannis Demiris},
title = {Unsupervised Learning of Complex Articulated Kinematic Structures
combining Motion and Skeleton Information},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
pages = {3138–3146},
month = {June},
year = {2015}


This work was supported in part by the EU FP7 project WYSIWYD under Grant 612139

For questions, contact