Evaluating the Effects of Convolutional Neural Network Committees

Fran Jurišić, Ivan Filković, Zoran Kalafatić

Abstract

Many high performing deep learning models for image classification put their base models in a committee as a final step to gain competitive edge. In this paper we focus on that aspect, analyzing how committee size and makeup of models trained with different preprocessing methods impact final performance. Working with two datasets, representing both rigid and non-rigid object classification in German Traffic Sign Recognition Benchmark (GTSRB) and CIFAR-10, and two preprocessing methods in addition to original images, we report performance improvements and compare them. Our experiments cover committees trained on just one dataset variation as well as hybrid ones, unreliability of small committees of low error models and performance metrics specific to the way committees are built. We point out some guidelines to predict committee behavior and good approaches to analyze their impact and limitations.

References

  1. Ciresan, D. C., Meier, U., Masci, J., and Schmidhuber, J. (2011). A committee of neural networks for traffic sign classification. InIJCNN, pages 1918-1921. IEEE.
  2. Ciresan, D. C., Meier, U., and Schmidhuber, J. (2012). Multi-column deep neural networks for image classification. InCVPR, pages 3642-3649. IEEE.
  3. Jia, Y., Shelhamer, E., Donahue, J., Karayev, S., Long, J., Girshick, R. B., Guadarrama, S., and Darrell, T. (2014). Caffe: Convolutional architecture for fast feature embedding. CoRR, abs/1408.5093.
  4. Jin, J., Fu, K., and Zhang, C. (2014). Traffic sign recognition with hinge loss trained convolutional neural networks. Intelligent Transportation Systems, IEEE Transactions on, PP(99):1-10.
  5. Krizhevsky, A. (2009). Learning multiple layers of features from tiny images.
  6. Krizhevsky, A. (2014). One weird trick for parallelizing convolutional neural networks. CoRR, abs/1404.5997.
  7. Krizhevsky, A., Sutskever, I., and Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. In Advances in Neural Information Processing Systems 25, pages 1097-1105. Curran Associates, Inc.
  8. Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M. S., Berg, A. C., and Fei-Fei, L. (2014). Imagenet large scale visual recognition challenge. CoRR, abs/1409.0575.
  9. Simonyan, K. and Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. CoRR.
  10. Stallkamp, J., Schlipsing, M., Salmen, J., and Igel, C. (2011). The German Traffic Sign Recognition Benchmark: A multi-class classification competition. In IEEE International Joint Conference on Neural Networks, pages 1453-1460.
  11. Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., and Rabinovich, A. (2015). Going deeper with convolutions. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR).
  12. Zeiler, M. D. and Fergus, R. (2014). Visualizing and understanding convolutional networks. In Computer Vision - ECCV 2014 - 13th European Conference, Zurich, pages 818-833.
Download


Paper Citation


in Harvard Style

Jurišić F., Filković I. and Kalafatić Z. (2016). Evaluating the Effects of Convolutional Neural Network Committees . In Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP, (VISIGRAPP 2016) ISBN 978-989-758-175-5, pages 560-565. DOI: 10.5220/0005719305600565


in Bibtex Style

@conference{visapp16,
author={Fran Jurišić and Ivan Filković and Zoran Kalafatić},
title={Evaluating the Effects of Convolutional Neural Network Committees},
booktitle={Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP, (VISIGRAPP 2016)},
year={2016},
pages={560-565},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0005719305600565},
isbn={978-989-758-175-5},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 4: VISAPP, (VISIGRAPP 2016)
TI - Evaluating the Effects of Convolutional Neural Network Committees
SN - 978-989-758-175-5
AU - Jurišić F.
AU - Filković I.
AU - Kalafatić Z.
PY - 2016
SP - 560
EP - 565
DO - 10.5220/0005719305600565