Boosted Convolutional Neural Networks

Mohammad Moghimi, Serge Belongie, Mohammad Saberian, Jian Yang, Nuno Vasconcelos and Li-Jia Li

Abstract

In this work, we propose a new algorithm for boosting Deep Convolutional Neural Networks (BoostCNN) to combine the merits of boosting and these networks. To learn this new model, we propose a novel algorithm to incorporate boosting weights into the deep learning architecture based on least square objective function. We also show that it is possible to use networks of different structures within the proposed boosting framework and BoostCNN is able to select the best network structure in each iteration. This not only results in superior performance but also reduces the required manual effort for finding the right network structure. Experiments show that the proposed method is able to achieve state-of-the-art performance on several fine-grained classification tasks such as bird, car, and aircraft classification.

Session

Posters 1

Files

PDF iconExtended Abstract (PDF, 91K)
PDF iconPaper (PDF, 321K)

DOI

10.5244/C.30.24
https://dx.doi.org/10.5244/C.30.24

Citation

Mohammad Moghimi, Serge Belongie, Mohammad Saberian, Jian Yang, Nuno Vasconcelos and Li-Jia Li. Boosted Convolutional Neural Networks. In Richard C. Wilson, Edwin R. Hancock and William A. P. Smith, editors, Proceedings of the British Machine Vision Conference (BMVC), pages 24.1-24.13. BMVA Press, September 2016.

Bibtex

        @inproceedings{BMVC2016_24,
        	title={Boosted Convolutional Neural Networks},
        	author={Mohammad Moghimi, Serge Belongie, Mohammad Saberian, Jian Yang, Nuno Vasconcelos and Li-Jia Li},
        	year={2016},
        	month={September},
        	pages={24.1-24.13},
        	articleno={24},
        	numpages={13},
        	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.30.24},
        	isbn={1-901725-59-6},
        	url={https://dx.doi.org/10.5244/C.30.24}
        }