Deep Part-Based Generative Shape Model with Latent Variables

Alexander Kirillov, Mikhail Gavrikov, Ekaterina Lobacheva, Anton Osokin and Dmitry Vetrov

Abstract

The Shape Boltzmann Machine (SBM) and its multilabel version MSBM have been recently introduced as deep generative models that capture the variations of an object shape. While being more flexible MSBM requires datasets with labeled parts of the objects for training. In the paper we present an algorithm for training MSBM using binary masks of objects and the seeds which approximately correspond to the locations of objects parts. The latter can be obtained from part-based detectors in an unsupervised manner. We derive a latent variable model and an EM-like training procedure for adjusting the weights of MSBM using a deep learning framework. We show that the model trained by our method outperforms SBM in the tasks related to binary shapes and is very close to the original MSBM in terms of quality of multilabel shapes.

Session

Posters 2

Files

Extended Abstract (PDF, 254K)
Paper (PDF, 1M)
Supplemental Materials (ZIP, 950K)

DOI

10.5244/C.30.88
https://dx.doi.org/10.5244/C.30.88

Citation

Bibtex

        @inproceedings{BMVC2016_88,
        	title={Deep Part-Based Generative Shape Model with Latent Variables},
        	author={Alexander Kirillov, Mikhail Gavrikov, Ekaterina Lobacheva, Anton Osokin and Dmitry Vetrov},
        	year={2016},
        	month={September},
        	pages={88.1-88.12},
        	articleno={88},
        	numpages={12},
        	booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
        	publisher={BMVA Press},
        	editor={Richard C. Wilson, Edwin R. Hancock and William A. P. Smith},
        	doi={10.5244/C.30.88},
        	isbn={1-901725-59-6},
        	url={https://dx.doi.org/10.5244/C.30.88}
        }