Abstract
The importance of the deep neural network lies in the fact that it opened the door to the complicated non-linear model and systematic approach for the hierarchical processing of knowledge.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
LeCun, Y., et al., “Handwritten digit recognition with a back-propagation network,” In Proc. Advances in Neural Information Processing Systems, 396–404 (1990).
- 2.
Krizhevsky, Alex, “ImageNet Classification with Deep Convolutional Neural Networks,” 17 November 2013.
- 3.
The representative methods include SIFT, HoG, Textons, Spin image, RIFT, and GLOH.
- 4.
It is often explained using the local receptive filed and shared weights from the perspective of the ordinary neural network. However, they would not be helpful for beginners. This book does not insist its relationship with the ordinary neural network and explains it as a type of digital filter.
- 5.
Also called kernels.
- 6.
deeplearning.stanford.edu/wiki/images/6/6c/Convolution_schematic.gif
- 7.
The designer decides how many elements to stride for each operation. It can be greater than one if the filter is larger.
- 8.
Sometimes the activation function is omitted depending on the problem.
- 9.
- 10.
loadMNISTImages and loadMNISTLabels functions are from github.com/amaas/stanford_dl_ex/tree/master/common.
- 11.
This size is valid only for this particular example. It varies depending on how the convolution filter is applied.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2017 Phil Kim
About this chapter
Cite this chapter
Kim, P. (2017). Convolutional Neural Network. In: MATLAB Deep Learning. Apress, Berkeley, CA. https://doi.org/10.1007/978-1-4842-2845-6_6
Download citation
DOI: https://doi.org/10.1007/978-1-4842-2845-6_6
Published:
Publisher Name: Apress, Berkeley, CA
Print ISBN: 978-1-4842-2844-9
Online ISBN: 978-1-4842-2845-6
eBook Packages: Professional and Applied ComputingApress Access BooksProfessional and Applied Computing (R0)