Attribute Recognition from Adaptive Parts

Luwei Yang, Ligeng Zhu, Yichen Wei, Shuang Liang and Ping Tan

Abstract

Previous part-based attribute recognition approaches perform part detection and attribute recognition in separate steps. The parts are not optimized for attribute recognition and therefore could be sub-optimal. We present an end-to-end deep learning approach to overcome the limitation. It generates object parts from key points and perform attribute recognition accordingly, allowing adaptive spatial transform of the parts. Both key point estimation and attribute recognition are learnt jointly in a multi-task setting. Extensive experiments on two datasets verify the efficacy of the proposed end-to-end approach.

Session

Posters 2

Files

PDF iconExtended Abstract (PDF, 2M)
PDF iconPaper (PDF, 7M)

DOI

10.5244/C.30.81
https://dx.doi.org/10.5244/C.30.81

Citation

Luwei Yang, Ligeng Zhu, Yichen Wei, Shuang Liang and Ping Tan. Attribute Recognition from Adaptive Parts. In Richard C. Wilson, Edwin R. Hancock and William A. P. Smith, editors, Proceedings of the British Machine Vision Conference (BMVC), pages 81.1-81.11. BMVA Press, September 2016.

Bibtex

        @inproceedings{BMVC2016_81,
        	title={Attribute Recognition from Adaptive Parts},
        	author={Luwei Yang, Ligeng Zhu, Yichen Wei, Shuang Liang and Ping Tan},
        	year={2016},
        	month={September},
        	pages={81.1-81.11},
        	articleno={81},
        	numpages={11},
        	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.30.81},
        	isbn={1-901725-59-6},
        	url={https://dx.doi.org/10.5244/C.30.81}
        }