Training method and apparatus for neural network for image recognition

• US 10,296,813 B2
• Filed: 09/01/2016
• Issued: 05/21/2019
• Est. Priority Date: 09/02/2015
• Status: Active Grant

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Chen, Li, Wang, Song, Fan, Wei, Sun, Jun, Satoshi, Naoi
• Primary Examiner(s): Grant, II, Jerome
• Application Number:
  US15/254,249
• Publication Number:
  US 20170061246A1
• Time in Patent Office: 992 Days
• Field of Search: 382156
• CPC Class Codes: G06F 18/214 Generating training patterns; Bootstrap methods, e.g. bagging or boosting
• CPC Class Codes: G06F 18/28 Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06T 2207/10004 Still image; Photographic image
• CPC Class Codes: G06T 2207/20084 Artificial neural networks [ANN]
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: G06V 10/772 Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries
• CPC Class Codes: G06V 10/774 Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting

Abstract

A training method and a training apparatus for a neutral network for image recognition are provided. The method includes: representing a sample image as a point set in a high-dimensional space, a size of the high-dimensional space being a size of space domain of the sample image multiplied by a size of intensity domain of the sample image; generating a first random perturbation matrix having a same size as the high-dimensional space; smoothing the first random perturbation matrix; perturbing the point set in the high-dimensional space using the smoothed first random perturbation matrix to obtain a perturbed point set; and training the neutral network using the perturbed point set as a new sample. With the training method and the training apparatus, classification performance of a conventional convolutional neural network is improved, thereby generating more training samples, reducing influence of overfitting, and enhancing generalization performance of the convolutional neural network.

Claims

1. A training method for a neural network for image recognition by a computing apparatus, the method comprising:

by the computing apparatus,

representing a sample image as a point set in a high-dimensional space, a space size of the high-dimensional space being a domain size of a space domain of the sample image multiplied by an intensity size of an intensity domain of the sample image;

generating a first random perturbation matrix to perturb a space domain and/or a intensity domain in a same space size as the high-dimensional space;

filtering information of the generated first random perturbation matrix to smooth the generated first random perturbation matrix;

perturbing the point set in the high-dimensional space using the smoothed first random perturbation matrix to obtain a perturbed point set; and

training the neural network using the perturbed point set as a new sample image for the image recognition.

2. The training method according to claim 1, wherein

the domain size of the space domain is a width of the sample image multiplied by a height of the sample image, and

the intensity size of the intensity domain is D1×

D2×

. . . ×

Dn, where n is a dimension of the intensity domain, D1, D2, . . . , Dn are sizes of respective dimensions.

3. The training method according to claim 1, wherein the training the neural network using the perturbed point set as a new sample comprises:

projecting the perturbed point set into one of a two-dimensional and three-dimensional sub-space of the high-dimensional space to obtain a projected image, and training the neural network using the projected image as a new sample.

4. The training method according to claim 1, wherein the first random perturbation matrix is generated using a given probability distribution, the given probability distribution being one of a Gaussian distribution and a uniform distribution.

5. The training method according to claim 1, wherein the first random perturbation matrix is smoothed using a low-pass filter having a same dimension as the first random perturbation matrix, the low-pass filter being one of a Gaussian filter, a mean filter and a middle value filter.

6. The training method according to claim 1, wherein the perturbing the point set in the high-dimensional space using the smoothed first random perturbation matrix to obtain a perturbed point set comprises:

adding a value of a corresponding element in the smoothed first random perturbation matrix to each point in the point set in the high-dimensional space.

7. The training method according to claim 1, by the computing apparatus further:

generating a second random perturbation matrix;

smoothing the second random perturbation matrix; and

perturbing, using the smoothed second random perturbation matrix, feature maps output by a convolution layer in training the neural network,

wherein a matrix size of the second random perturbation matrix is a length of a feature map multiplied by a width of the feature map multiplied by a number of the feature maps.

8. The training method according to claim 7, wherein the perturbing, using the smoothed second random perturbation matrix, feature maps outputted by a convolution layer in training the neural network comprises:

adding a corresponding weighted value of the second random perturbation matrix to a value of the feature map, wherein a weighting coefficient is larger if an absolute value of the feature map is larger.

19. A non-transitory computer readable medium storing a training method for a neural network for image recognition as recited in claim 1.

9. A training method for a neural network for image recognition by a computing apparatus, the method comprising:

by the computing apparatus,

transforming a sample image to obtain a transform domain representation of the sample image;

representing a transformed sample image as a point set in a high-dimensional space, a space size of the high-dimensional space being a domain size of a space domain of the sample image multiplied by an intensity size of an intensity domain of the sample image;

generating a random perturbation matrix to perturb a space domain and/or a intensity domain in a same space size as the high-dimensional space;

filtering information of the generated random perturbation matrix to smooth the generated random perturbation matrix;

perturbing the point set in the high-dimensional space using the smoothed random perturbation matrix to obtain a perturbed point set;

projecting the perturbed point set into a transform domain sub-space of the high-dimensional space to obtain a transform domain representation of a perturbed sample image;

inversely transforming the transform domain representation of the perturbed sample image to obtain a perturbed image; and

training the neural network using the perturbed image as a new sample image for the image recognition.

10. The training method according to claim 9, wherein the transform is one of a Discrete Fourier Transform and a Discrete Cosine Transform.

11. A training apparatus for a neural network for image recognition, comprising:

a memory; and

a processor coupled to the memory to,

a sample image as a point set in a high-dimensional space, a space size of the high-dimensional space being a domain size of a space domain of the sample image multiplied by an intensity size of an intensity domain of the sample image;

generate a first random perturbation matrix to perturb a space domain and/or a intensity domain in a same space size as the high-dimensional space;

filtering information of the generated first random perturbation matrix to smooth the generated first random perturbation matrix;

perturb the point set in the high-dimensional space using the smoothed first random perturbation matrix to obtain a perturbed point set; and

train the neural network using the perturbed point set as a new sample image for the image recognition.

12. The training apparatus according to claim 11, wherein the domain size of the space domain is a width of the sample image multiplied by a height of the sample image, and the intensity size of the intensity domain is D1×

D2×

. . . ×

Dn, where n is a dimension of the intensity domain, D1, D2, . . . , Dn are sizes of respective dimensions.

13. The training apparatus according to claim 11, wherein the processor is to:

project the perturbed point set into one of a two-dimensional and three-dimensional sub-space of the high-dimensional space to obtain a projected image, and train the neural network using the projected image as the new sample image.

14. The training apparatus according to claim 11, wherein the first random perturbation matrix is generated using a given probability distribution, the given probability distribution being one of a Gaussian distribution and a uniform distribution.

15. The training apparatus according to claim 11, wherein the processor is to filter the first random perturbation matrix using a low-pass filter having a same dimension as the first random perturbation matrix, the low-pass filter being one of a Gaussian filter, a mean filter and a middle value filter.

16. The training apparatus according to claim 11, wherein the processor is to:

add a value of a corresponding element in the smoothed first random perturbation matrix to each point in the point set in the high-dimensional space.

17. The training apparatus according to claim 11, wherein the processor is to:

generate a second random perturbation matrix;

filter information of the generated second random perturbation matrix to smooth the generated second random perturbation matrix; and

perturb, using the smoothed second random perturbation matrix, feature maps output by a convolution layer in training the neural network,

wherein a matrix size of the second random perturbation matrix is a length of a feature map multiplied by a width of the feature map multiplied by a number of the feature maps.

18. The training apparatus according to claim 17, wherein the processor is to:

add a corresponding weighted value of the second random perturbation matrix to a value of the feature map, wherein a weighting coefficient is larger if an absolute value of the feature map is larger.

20. A training apparatus for a neural network for image recognition, comprising:

a computer, comprising;

a high-dimensional space representing unit configured to represent a sample image as a point set in a high-dimensional space, a space size of the high-dimensional space being a domain size of a space domain of the sample image multiplied by an intensity size of an intensity domain of the sample image;

a random perturbation matrix generating unit configured to generate a random perturbation matrix to perturb a space domain and/or a intensity domain in having a same space size as the high-dimensional space;

a smoothing unit configured to filter information of the generated random perturbation matrix to smooth the generated random perturbation matrix;

a perturbing unit configured to perturb the point set in the high-dimensional space using the smoothed random perturbation matrix to obtain a perturbed point set; and

a training sample determining unit configured to train the neural network using the perturbed point set as a new sample image for the image recognition.