Rapid Exact Signal Scanning With Deep Convolutional Neural Networks
作者： Markus Thom , Franz Gritschneder
关键词: Signal 、 Algorithm 、 Massively parallel 、 Convolutional neural network 、 Computational complexity theory 、 Artificial neural network 、 Sliding window protocol 、 Convolution 、 Computer science 、 Signal processing
摘要: A rigorous formulation of the dynamics a signal processing scheme aimed at dense scanning without any loss in accuracy is introduced and analyzed. Related methods proposed recent past lack satisfactory analysis whether they actually fulfill exactness constraints. This improved through an exact characterization requirements for sound sliding window approach. The tools developed this paper are especially beneficial if Convolutional Neural Networks employed, but can also be used as more general framework to validate related approaches scanning. theory helps eliminate redundant computations renders special case treatment unnecessary, resulting dramatic boost efficiency particularly on massively parallel processors. demonstrated both theoretically computational complexity empirically modern
harvard.edu
arxiv.org
sci-hub.se 参考文章(27)
Hongsheng Li, Xiaogang Wang, Rui Zhao, Highly Efficient Forward and Backward Propagation of Convolutional Neural Networks for Pixelwise Classification arXiv: Computer Vision and Pattern Recognition. ,(2014)
Christopher M. Bishop, Neural networks for pattern recognition ,(1995)
Terence D. Sanger, Optimal Unsupervised Learning in Feedforward Neural Networks Massachusetts Institute of Technology. ,(1989)
Evan Shelhamer, John Tran, Jonathan Cohen, Sharan Chetlur, Philippe Vandermersch, Cliff Woolley, Bryan Catanzaro, cuDNN: Efficient Primitives for Deep Learning arXiv: Neural and Evolutionary Computing. ,(2014)
Chao Dong, Chen Change Loy, Kaiming He, Xiaoou Tang, Image Super-Resolution Using Deep Convolutional Networks IEEE Transactions on Pattern Analysis and Machine Intelligence. ,vol. 38, pp. 295- 307 ,(2016) , 10.1109/TPAMI.2015.2439281
L. Dagum, R. Menon, OpenMP: an industry standard API for shared-memory programming computational science and engineering. ,vol. 5, pp. 46- 55 ,(1998) , 10.1109/99.660313
Clement Farabet, Camille Couprie, Laurent Najman, Yann LeCun, Learning Hierarchical Features for Scene Labeling IEEE Transactions on Pattern Analysis and Machine Intelligence. ,vol. 35, pp. 1915- 1929 ,(2013) , 10.1109/TPAMI.2012.231
H. Neudecker, Some Theorems on Matrix Differentiation with Special Reference to Kronecker Matrix Products Journal of the American Statistical Association. ,vol. 64, pp. 953- 963 ,(1969) , 10.1080/01621459.1969.10501027
Alessandro Giusti, Dan C. Ciresan, Jonathan Masci, Luca M. Gambardella, Jurgen Schmidhuber, Fast image scanning with deep max-pooling convolutional neural networks international conference on image processing. pp. 4034- 4038 ,(2013) , 10.1109/ICIP.2013.6738831
M Sai Praneeth, Xudong Peng, Alice Li, Shahrzad Hosseini Vajargah, Going deeper with convolutions computer vision and pattern recognition. pp. 1- 9 ,(2015) , 10.1109/CVPR.2015.7298594