Processing

Please wait...

Settings

Settings

Goto Application

1. WO2018145604 - SAMPLE SELECTION METHOD, APPARATUS AND SERVER

Publication Number WO/2018/145604
Publication Date 16.08.2018
International Application No. PCT/CN2018/075114
International Filing Date 02.02.2018
IPC
G06K 9/62 2006.1
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
CPC
G06F 16/00
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
16Information retrieval; Database structures therefor; File system structures therefor
G06K 9/62
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
G06K 9/6215
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
6201Matching; Proximity measures
6215Proximity measures, i.e. similarity or distance measures
G06K 9/6254
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
6217Design or setup of recognition systems and techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
6253User interactive design
6254Interactive pattern learning with a human teacher
G06K 9/6255
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
6217Design or setup of recognition systems and techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
6255Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries, e.g. user dictionaries
G06K 9/6256
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
6217Design or setup of recognition systems and techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
6256Obtaining sets of training patterns; Bootstrap methods, e.g. bagging, boosting
Applicants
  • 南京航空航天大学 NANJING UNIVERSITY OF AERONAUTICS AND ASTRONAUTICS [CN]/[CN]
  • 腾讯科技(深圳)有限公司 TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED [CN]/[CN]
Inventors
  • 黄圣君 HUANG, Shengjun
  • 高能能 GAO, Nengneng
  • 袁坤 YUAN, Kun
  • 陈伟 CHEN, Wei
  • 王迪 WANG, Di
Agents
  • 北京三高永信知识产权代理有限责任公司 BEIJING SAN GAO YONG XIN INTELLECTUAL PROPERTY AGENCY CO., LTD.
Priority Data
201710069595.X08.02.2017CN
Publication Language Chinese (zh)
Filing Language Chinese (ZH)
Designated States
Title
(EN) SAMPLE SELECTION METHOD, APPARATUS AND SERVER
(FR) PROCÉDÉ, APPAREIL ET SERVEUR DE SÉLECTION D'ÉCHANTILLONS
(ZH) 样本选择方法、装置及服务器
Abstract
(EN) A sample selection method, an apparatus and a server, which relate to the technical field of metric learning. Said method comprises: selecting n sample pairs from an unlabeled sample set, each sample pair comprising two samples, and each sample comprising p modes of data (101); for each sample pair, calculating the partial degree of similarity between each mode of data of a sample which the sample pair comprises and each mode of data of the other sample to obtain the p×p partial degree of similarity (102); calculating the overall degree of similarity between the two samples which the sample pair comprise according to the p×p partial degree of similarity (103); obtaining the degree of difference between the p×p partial degree of similarity and the overall degree of similarity (104); selecting a sample pair from the n sample pairs which meets a pre-set condition to act as a training sample (105). By means of selecting a high-quality training sample to train a metric model, said method, apparatus and server may use fewer training samples to train metric models having higher accuracy.
(FR) L'invention concerne un procédé de sélection d'échantillons, un appareil et un serveur qui se rapportent au domaine technique de l'apprentissage métrique. Ledit procédé consiste à : sélectionner n paires d'échantillons à partir d'un ensemble d'échantillons non étiquetés, chaque paire d'échantillons comprenant deux échantillons, et chaque échantillon comprenant p modes de données (101) ; pour chaque paire d'échantillons, calculer le degré partiel de similarité entre chaque mode de données d'un échantillon inclus dans la paire d'échantillons et chaque mode de données de l'autre échantillon afin d'obtenir le degré partiel de similarité p×p (102) ; calculer le degré global de similarité entre les deux échantillons inclus dans la paire d'échantillons en fonction du degré partiel de similarité p×p (103) ; obtenir le degré de différence entre le degré partiel de similarité p×p et le degré global de similarité (104) ; et sélectionner une paire d'échantillons parmi les n paires d'échantillons qui remplit une condition prédéfinie pour servir d'échantillon d'apprentissage (105). En sélectionnant un échantillon d'apprentissage de haute qualité pour former un modèle métrique, ledit procédé, ledit appareil et ledit serveur peuvent utiliser un nombre réduit d'échantillons d'apprentissage pour former des modèles métriques ayant une précision plus élevée.
(ZH) 一种样本选择方法、装置及服务器,属于度量学习技术领域。所述方法包括:从未标注样本集中选取n组样本对,每一组样本对包括两个样本,每一个样本包括p种模态的数据(101);对于每一组样本对,计算样本对包括的一个样本的每一种模态的数据和另一个样本的每一种模态的数据之间的部分相似度,得到p×p个部分相似度(102);根据p×p个部分相似度计算样本对包括的两个样本之间的整体相似度(103);获取p×p个部分相似度与整体相似度之间的差异程度(104);从n组样本对中选择符合预设条件的样本对作为训练样本(105)。所述方法、装置及服务器通过选择高质量的训练样本训练度量模型,能够用更少的训练样本训练出更高精度的度量模型。
Related patent documents
Latest bibliographic data on file with the International Bureau