Geodesic
Neural network training acceleration using NVIDIA CUDA technology for image recognition
Journal of Samara State Technical University, Ser. Physical and Mathematical Sciences, Tome 126 (2012) no. 1, pp. 183-191 Cet article a éte moissonné depuis la source Math-Net.Ru

Voir la notice de l'article

In this paper, an implementation of neural network trained by algorithm based on Levenberg-Marquardt method is presented. Training of neural network increased by almost 9 times using NVIDIA CUDA technology. Implemented neural network is used for the recognition of noised images.
Keywords: Image recognition, neural networks, graphics processing unit, CUDA.
Mots-clés : Levenberg-Marquardt method 
@article{VSGTU_2012_126_1_a18,
     author = {A. A. Fertsev},
     title = {Neural network training acceleration using {NVIDIA} {CUDA} technology for image recognition},
     journal = {Journal of Samara State Technical University, Ser. Physical and Mathematical Sciences},
     pages = {183--191},
     year = {2012},
     volume = {126},
     number = {1},
     language = {ru},
     url = {http://geodesic.mathdoc.fr/item/VSGTU_2012_126_1_a18/}
}
TY  - JOUR
AU  - A. A. Fertsev
TI  - Neural network training acceleration using NVIDIA CUDA technology for image recognition
JO  - Journal of Samara State Technical University, Ser. Physical and Mathematical Sciences
PY  - 2012
SP  - 183
EP  - 191
VL  - 126
IS  - 1
UR  - http://geodesic.mathdoc.fr/item/VSGTU_2012_126_1_a18/
LA  - ru
ID  - VSGTU_2012_126_1_a18
ER  - 
%0 Journal Article
%A A. A. Fertsev
%T Neural network training acceleration using NVIDIA CUDA technology for image recognition
%J Journal of Samara State Technical University, Ser. Physical and Mathematical Sciences
%D 2012
%P 183-191
%V 126
%N 1
%U http://geodesic.mathdoc.fr/item/VSGTU_2012_126_1_a18/
%G ru
%F VSGTU_2012_126_1_a18
A. A. Fertsev. Neural network training acceleration using NVIDIA CUDA technology for image recognition. Journal of Samara State Technical University, Ser. Physical and Mathematical Sciences, Tome 126 (2012) no. 1, pp. 183-191. http://geodesic.mathdoc.fr/item/VSGTU_2012_126_1_a18/

[1] Mashor M. Y., Sulaiman S. N., “Recognition of Noisy Numerals using Neural Network”, IJCIM, 9:3 (2001), 158–164

[2] Boureau Y.-L., Bach F., LeCun Y., Ponce J., “Learning mid-level features for recognition”, IEEE Conference on Computer Vision and Pattern Recognition, 2010, 2559–2566 | DOI

[3] Hagan M. T., Menhaj M., “Training feedforward networks with the Marquardt algorithm”, IEEE Transactions on Neural Networks, 5:6 (1994), 989–993 | DOI

[4] Wilamowski B. M., Chen Y., Malinowski A., “Efficient algorithm for training neural networks with one hidden layer”, International Joint Conference on Neural Networks (IJCNN '99), v. 3, 1999, 1725–1728 | DOI

[5] Marquardt D., “An Algorithm for Least-Squares Estimation of Nonlinear Parameters”, SIAM J. Appl. Math., 11:2 (1963), 431–441 | DOI | MR | Zbl

[6] David J. C. MacKay, “A practical Bayesian framework for backpropagation networks”, Neural Computation, 4:3 (1992), 448–472 | DOI

[7] Poland J., On the Robustness of update strategies for the Bayesian hyperparameter alpha, available on: , 2001 http://www-alg.ist.hokudai.ac.jp/~jan/alpha.pdf

[8] Nguyen D., Widrow B., “Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights”, International Joint Conference on Neural Networks, v. 3, 1990, 21–26 | DOI

[9] Izotov P. Yu., Sukhanov S. V., Golovashkin D. L., “Technology of implementation of neural network algorithm in CUDA environment at the example of handwritten digits recognition”, Komp'yuternaya optika, 34:2 (2010), 243–252

[10] Gonzalez R. C., Woods R. E., Digital Image Processing, Addison-Wesley Publishing Company, Boston, MA, 1992, 528 pp.