Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
IMAGE RECOGNITION DEVICE, IMAGE RECOGNITION METHOD, AND IMAGE RECOGNITION PROGRAM
Document Type and Number:
WIPO Patent Application WO/2018/207334
Kind Code:
A1
Abstract:
The image recognition device according to the present invention is provided with: a parameter calculation means which, by using at least one or more first images having a predetermined attribute that is common to subjects to be authenticated, calculates one or more parameters for extracting, from a second image, a feature indicating that the second image is likely to have the predetermined attribute, or unlikely to have the predetermined attribute, wherein it is not known whether or not the second image has the predetermined attribute; a feature extraction means which extracts a feature from an input image using the one or more parameters; and a determination means which determines whether or not the second image is a subject to be authenticated, on the basis of at least said feature extracted from the second image. The feature extraction means extracts a feature using a neural network that is obtained by connecting together calculation elements having predetermined parameters, and that includes two or more layers including input and output layers, each layer comprising calculation elements, wherein at least one layer comprises fewer calculation elements than the input layer, to which image information is input.

Inventors:
AKIYAMA TATSUO (JP)
Application Number:
PCT/JP2017/017985
Publication Date:
November 15, 2018
Filing Date:
May 12, 2017
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
NEC CORP (JP)
International Classes:
G06T7/00
Domestic Patent References:
WO2015008567A12015-01-22
Foreign References:
JP2017004350A2017-01-05
JP2014203135A2014-10-27
JPH09161054A1997-06-20
Other References:
SHO SONODA ET AL.: "Transportation aspect of infinitely deep denoising autoencoder", IEICE TECHNICAL REPORT, vol. 116, no. 300, 9 November 2016 (2016-11-09), pages 297 - 304
Attorney, Agent or Firm:
IWAKABE Fuyuki et al. (JP)
Download PDF: