Learning method and recording medium

• US 9,779,354 B2
• Filed: 02/25/2016
• Issued: 10/03/2017
• Est. Priority Date: 03/06/2015
• Status: Active Grant

• Current Assignee: Panasonic Intellectual Property Management Co., Ltd. (Panasonic Holdings Corporation)
• Original Assignee: Panasonic Intellectual Property Management Co., Ltd. (Panasonic Holdings Corporation)
• Inventors: Hagawa, Reiko, Tsukizawa, Sotaro, Ishii, Yasunori
• Primary Examiner(s): Seth, Manav
• Application Number:
  US15/053,642
• Publication Number:
  US 20160260014A1
• Time in Patent Office: 586 Days
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06F 18/24317 Piecewise classification, i.e. whereby each classification requires several discriminant rules
• CPC Class Codes: G06F 18/24323 Tree-organised classifiers
• CPC Class Codes: G06K 7/1482 using fuzzy logic or natural solvers, such as neural networks, genetic algorithms and simulated annealing
• CPC Class Codes: G06N 3/02 Neural networks
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06V 10/454 Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 10/82 using neural networks

Abstract

Learning method includes performing a first process in which a coarse class classifier configured with a first neural network is made to classify a plurality of images given as a set of images each attached with a label indicating a detailed class into a plurality of coarse classes including a plurality of detailed classes and is then made to learn a first feature that is a feature common in each of the coarse classes, and performing a second process in which a detailed class classifier, configured with a second neural network that is the same in terms of layers other than the final layer as but different in terms of the final layer from the first neural network made to perform the learning in the first process, is made to classify the set of images into detailed classes and learn a second feature of each detailed class.

Claims

1. A method of learning a classifier for classifying an image, comprising:

performing a first process, in which a coarse class classifier, configured with a first neural network, classifies a plurality of images, comprising a set of images each having attached a label indicating a detailed class into a plurality of coarse classes including a plurality of detailed classes and then learns a first feature that is a common to each of the coarse classes; and

performing a second process in which a detailed class classifier classifies the set of images into a plurality of detailed classes and then learns a second feature that is common to each of the detailed classes, the detailed class classifier configured with a second neural network that is the same in terms of layers other than a final layer, and is different in terms of the final layer, from the first neural network that performs the learning in the first process,

wherein at least one of the first process and the second process is performed by a processor.

2. The method according to claim 1, wherein

the first process includes a first design process and a first learning process, the first design process including designing the first neural network to include as many neurons in the final layer as there are coarse classes generated based on contents of the labels indicating the respective detailed classes, the first learning process including making the coarse class classifier, configured with the designed first neural network, acquire the set of images and learn the first feature,

the second process includes a second design process and a second learning process, the second design process including designing the second neural network such that the neurons in the final layer of the first neural network that performs the learning in the first process are changed by a number of neurons that is the same as the number of detailed classes, and the resultant neural network is employed as the second neural network, the second learning process including making the detailed class classifier, configured with the designed second neural network, acquire the set of images and learn the second feature.

3. The method according to claim 2, wherein, in the second design process, the second neural network is designed such that the neurons in the final layer of the first neural network that performs the learning in the first learning process are replaced with a number of neurons that is the same as the number of detailed classes and the resultant neural network is employed as the second neural network.

4. The method according to claim 2, wherein

each of the detailed class classifier and the coarse class classifier is formed using a third neural network,

wherein the method further comprises a third design process before the first process, the third design process including designing the third neural network such that a final layer of the third neural network includes a number of neurons that is equal to or greater than a sum of the number of coarse classes and the number of detailed classes,

wherein in the first design process, the first neural network is given by the third neural network, except that, in the final layer of the third neural network, a number of particular neurons that is equal to the number of coarse classes are selected and only the selected neurons are used to design the first neural network to include as many neurons in the final layer as there are coarse classes generated based on contents of the labels indicating the respective detailed classes, and

in the second design process, the second neural network is given by the third neural network, except that, in the final layer of the third neural network, a number of particular neurons, which are different from the neurons employed as the neurons in the final layer of the first neural network, and that is equal to the number of detailed classes, are used to design the second neural network obtained by changing the neurons in the final layer of the first neural network that performs the learning in the first process with a number of neurons that is equal to the number of detailed classes.

5. The method according to claim 2, wherein the second design process further includes setting a weight of the final layer of the second neural network to a random number, the weight indicating a coupling efficiency between the final layer of the second neural network and a layer immediately in front of the final layer of the second neural network.

6. The method according to claim 2, wherein the second design process further includes setting a second weight of the final layer of the second neural network based on a first weight of the final layer of the first neural network, the first weight indicating a coupling efficiency, obtained via the learning in the first learning process, between the final layer of the first neural network and the layer immediately in front of the final layer of the first neural network, the second weight indicating a coupling efficiency between the final layer of the second neural network and the layer immediately in front of the final layer of the second neural network, the second weight being determined by proportionally dividing the first weight according to a relationship between the plurality of coarse classes and the plurality of detailed classes included in the plurality of coarse classes.

7. The method according to claim 1, further comprising storing, as the classifier, the detailed class classifier that learns the second feature in the second process.

8. The method according to claim 1, wherein the first process further includes generating the plurality of coarse classes based on contents of the labels attached to respective images in the plurality of images.

9. The method according to claim 1, wherein the first process further includes generating the plurality of coarse classes based on a similarity of each image in the plurality of images.

10. The method according to claim 1, further comprising

performing a third process, in which a third coarse class classifier, configured with a third neural network, classifies the set of images, into a plurality of third coarse classes including a plurality of detailed classes and then learns a third feature that is common to each third coarse class of the plurality of third coarse classes,

wherein the first process includes a first learning process, in which the detailed class classifier, configured with the first neural network that is the same in terms of layers other than the final layer, and is different in terms of the final layer from the third neural network that performs the learning in the third process, classifies the set of images into the plurality of coarse classes and learns the first feature of each detailed class,

and wherein each coarse class in the plurality of coarse classes includes a plurality of third coarse classes and includes a plurality of detailed classes included in the plurality of third coarse classes.

11. A non-transitory computer-readable recording medium storing a program for learning a classifier that classifies an image, the program causing a computer to execute:

performing a first process, in which a coarse class classifier, configured with a first neural network, classifies a plurality of images, comprising a set of images each having attached a label indicating a detailed class, into a plurality of coarse classes including a plurality of detailed classes and then learns a first feature that is a common to each of the coarse classes; and

performing a second process, in which a detailed class classifier classifies the set of images into a plurality of detailed classes and learns a second feature that is common to each of the detailed classes, the detailed class classifier configured with a second neural network that is the same in terms of layers other than a final layer, and is different in terms of the final layer, from the first neural network that performs the learning in the first process.