Processing

Please wait...

Settings

Settings

Goto Application

1. WO2020071143 - MULTI-DIMENSIONAL DATA VISUALIZATION DEVICE, METHOD, AND PROGRAM

Publication Number WO/2020/071143
Publication Date 09.04.2020
International Application No. PCT/JP2019/036886
International Filing Date 20.09.2019
IPC
G06F 16/26 2019.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
FELECTRIC DIGITAL DATA PROCESSING
16Information retrieval; Database structures therefor; File system structures therefor
20of structured data, e.g. relational data
26Visual data mining; Browsing structured data
A61B 5/05 2006.01
AHUMAN NECESSITIES
61MEDICAL OR VETERINARY SCIENCE; HYGIENE
BDIAGNOSIS; SURGERY; IDENTIFICATION
5Measuring for diagnostic purposes; Identification of persons
05Measuring for diagnosis by means of electric currents or magnetic fields
A61B 5/11 2006.01
AHUMAN NECESSITIES
61MEDICAL OR VETERINARY SCIENCE; HYGIENE
BDIAGNOSIS; SURGERY; IDENTIFICATION
5Measuring for diagnostic purposes; Identification of persons
103Measuring devices for testing the shape, pattern, size or movement of the body or parts thereof, for diagnostic purposes
11Measuring movement of the entire body or parts thereof, e.g. head or hand tremor or mobility of a limb
G06F 16/23 2019.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
FELECTRIC DIGITAL DATA PROCESSING
16Information retrieval; Database structures therefor; File system structures therefor
20of structured data, e.g. relational data
23Updating
G06F 16/248 2019.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
FELECTRIC DIGITAL DATA PROCESSING
16Information retrieval; Database structures therefor; File system structures therefor
20of structured data, e.g. relational data
24Querying
248Presentation of query results
CPC
A61B 5/05
AHUMAN NECESSITIES
61MEDICAL OR VETERINARY SCIENCE; HYGIENE
BDIAGNOSIS; SURGERY; IDENTIFICATION
5Detecting, measuring or recording for diagnostic purposes
05Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; ; Measuring using microwaves or radiowaves
A61B 5/11
AHUMAN NECESSITIES
61MEDICAL OR VETERINARY SCIENCE; HYGIENE
BDIAGNOSIS; SURGERY; IDENTIFICATION
5Detecting, measuring or recording for diagnostic purposes
103Detecting, measuring or recording devices for testing the shape, pattern, ; colour,; size or movement of the body or parts thereof, for diagnostic purposes
11Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
G06F 16/00
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
16Information retrieval; Database structures therefor; File system structures therefor
G06F 16/23
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
16Information retrieval; Database structures therefor; File system structures therefor
20of structured data, e.g. relational data
23Updating
G06F 16/248
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
16Information retrieval; Database structures therefor; File system structures therefor
20of structured data, e.g. relational data
24Querying
248Presentation of query results
G06F 16/26
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
16Information retrieval; Database structures therefor; File system structures therefor
20of structured data, e.g. relational data
26Visual data mining; Browsing structured data
Applicants
  • 日本電信電話株式会社 NIPPON TELEGRAPH AND TELEPHONE CORPORATION [JP]/[JP]
Inventors
  • 伊勢崎 隆司 ISEZAKI, Takashi
  • 青木 良輔 AOKI, Ryosuke
  • 渡部 智樹 WATANABE, Tomoki
  • 山田 智広 YAMADA, Tomohiro
Agents
  • 蔵田 昌俊 KURATA, Masatoshi
  • 野河 信久 NOGAWA, Nobuhisa
  • 峰 隆司 MINE, Takashi
  • 井上 正 INOUE, Tadashi
Priority Data
2018-18936304.10.2018JP
Publication Language Japanese (JA)
Filing Language Japanese (JA)
Designated States
Title
(EN) MULTI-DIMENSIONAL DATA VISUALIZATION DEVICE, METHOD, AND PROGRAM
(FR) DISPOSITIF DE VISUALISATION DE DONNÉES MULTIDIMENSIONNELLES, PROCÉDÉ ET PROGRAMME
(JA) 多次元データ可視化装置、方法およびプログラム
Abstract
(EN)
An embodiment of the present invention is provided with a projective transformation model which has a plurality of nodes that retain reference vectors having dimensions corresponding to multi-dimensional data, and which also has a projection table. The projection table represents the correspondence between the numbers of the nodes and the coordinates in a two-dimensional space to which the reference vector retained by each node is projected. Initially in a learning phase, positive-example and negative-example multi-dimensional input data are acquired, amplitude feature quantities of the respective multi-dimensional input data are calculated, and data of the respective amplitude feature quantities are learned for each sample as the reference vectors of the plurality of nodes. Next, a calculation is made of the Euclidean distance between coordinates when a node learned using the positive-example amplitude feature quantity data and a node learned using the negative-example amplitude feature quantity data have been projected to the two-dimensional space in accordance with the projection table, and coordinates in the projection table are updated such that the calculated Euclidean distance is at least equal to a threshold value.
(FR)
Selon un mode de réalisation, la présente invention concerne un modèle de transformation projectif qui a une pluralité de nœuds qui retiennent des vecteurs de référence ayant des dimensions correspondant à des données multidimensionnelles, et qui a également une table de projection. La table de projection représente la correspondance entre les nombres des nœuds et les coordonnées dans un espace bidimensionnel auquel le vecteur de référence retenu par chaque nœud est projeté. Initialement dans une phase d'apprentissage, des données d'entrée multidimensionnelles d'exemple positif et d'exemple négatif sont acquises, des quantités caractéristiques d'amplitude des données d'entrée multidimensionnelles respectives sont calculées, et des données des quantités caractéristiques d'amplitude respectives sont apprises pour chaque échantillon en tant que vecteurs de référence de la pluralité de nœuds. Ensuite, un calcul est effectué à partir de la distance euclidienne entre des coordonnées lorsqu'un nœud appris à l'aide des données de quantité caractéristique d'amplitude d'exemple positif et un nœud appris à l'aide des données de quantité caractéristique d'amplitude d'exemple négatif ont été projetés sur l'espace bidimensionnel conformément à la table de projection, et les coordonnées dans la table de projection sont mises à jour de telle sorte que la distance euclidienne calculée est au moins égale à une valeur seuil.
(JA)
この発明の一実施形態は、多次元データに対応する次元の参照ベクトルを保持する複数のノードと、射影テーブルとを有する射影変換モデルを備える。射影テーブルは、上記ノードの番号と、各ノードが保持する参照ベクトルの射影先となる二次元空間における座標との対応関係を表す。そして、先ず学習フェーズにおいて、正例および負例の各多次元入力データを取得してそれぞれその振幅特徴量を算出し、当該各振幅特徴量データをサンプルごとに上記複数のノードの参照ベクトルとして学習させる。次に、正例の振幅特徴量データにより学習されたノードと、負例の振幅特徴量データにより学習されたノードを、射影テーブルに従い二次元空間に射影したときの各座標のユークリッド距離を算出し、算出されたユークリッド距離が閾値以上となるように上記射影テーブルの座標を更新する。
Also published as
Latest bibliographic data on file with the International Bureau