US 11,755,907 B2
Feature identification device, feature identification method, and computer readable medium
Eri Kataoka, Tokyo (JP); Hiroshi Yokochi, Tokyo (JP); Mitsuhiro Matsumoto, Tokyo (JP); Mikihito Kanno, Tokyo (JP); Takamichi Koide, Tokyo (JP); and Yosuke Watanabe, Tokyo (JP)
Assigned to Mitsubishi Electric Corporation, Tokyo (JP)
Filed by Mitsubishi Electric Corporation, Tokyo (JP)
Filed on Jul. 29, 2021, as Appl. No. 17/388,283.
Application 17/388,283 is a continuation of application No. PCT/JP2019/046249, filed on Nov. 26, 2019.
Claims priority of application No. JP2019-056844 (JP), filed on Mar. 25, 2019.
Prior Publication US 2021/0357635 A1, Nov. 18, 2021
Int. Cl. G06N 3/08 (2023.01); G06V 30/414 (2022.01); G06F 18/22 (2023.01); G06F 18/241 (2023.01); G06N 3/045 (2023.01); G06V 30/40 (2022.01); G06V 30/19 (2022.01); G06V 10/82 (2022.01)
CPC G06N 3/08 (2013.01) [G06F 18/22 (2023.01); G06F 18/241 (2023.01); G06N 3/045 (2023.01); G06V 10/82 (2022.01); G06V 30/19173 (2022.01); G06V 30/40 (2022.01); G06V 30/414 (2022.01)] 8 Claims
OG exemplary drawing
 
1. A feature identification device comprising:
processing circuitry to:
treat each of a plurality of classifications as a target classification, and calculate a similarity between a reference feature value and a recognition feature value, the reference feature value being a feature value extracted from image data of the target classification, the recognition feature value being a feature value extracted from recognition target data, which is image data to be recognized;
calculate an influence on the similarity with regard to each partial image of the recognition target data by taking as input the similarity calculated with regard to each of the plurality of classifications and the recognition feature value; and
change the recognition feature value in accordance with the calculated influence.