Title:
マルチ関連ラベルを生成する方法及びシステム
Document Type and Number:
Japanese Patent JP6741357
Kind Code:
B2
Abstract:
A method for performing multi-label classification includes extracting a feature vector from an input vector including input data by a feature extractor, determining, by a label predictor, a relevant vector including relevant labels having relevant scores based on the feature vector, updating a binary masking vector by masking pre-selected labels having been selected in previous label selections, applying the updated binary masking vector to the relevant vector such that the relevant label vector is updated to exclude the pre-selected labels from the relevant labels, and selecting a relevant label from the updated relevant label vector based on the relevant scores of the updated relevant label vector.
Inventors:
Takaaki Hori
Tori Hori
Shinji Watanabe
Hershey, John
Harsham, Brett
Le Roux, Jonathan
Tori Hori
Shinji Watanabe
Hershey, John
Harsham, Brett
Le Roux, Jonathan
Application Number:
JP2019504995A
Publication Date:
August 19, 2020
Filing Date:
September 08, 2017
Export Citation:
Assignee:
Mitsubishi Electric Corporation
International Classes:
G06F16/906; G06F16/908; G06N20/00
Domestic Patent References:
| JP2016151805A | ||||
| JP2009151395A |
Other References:
Jiang Wang, 外5名,CNN-RNN: A Unified Framework for Multi-label Image Classification,IEEE Conference Proceedings,米国,2016年,Vol.2016 No.CVPR,pp.2285~2294
Attorney, Agent or Firm:
Michiharu Soga
Kajinami order
Kazuhiro Oyaku
Shunichi Ueda
Junichiro Yoshida
Kajinami order
Kazuhiro Oyaku
Shunichi Ueda
Junichiro Yoshida