TOP > 外国特許検索 > EVALUATION DEVICE, EVALUATION METHOD, PROGRAM, AND INFORMATION RECORDING MEDIUM

EVALUATION DEVICE, EVALUATION METHOD, PROGRAM, AND INFORMATION RECORDING MEDIUM

外国特許コード F190009886
整理番号 08697-WO
掲載日 2019年8月23日
出願国 世界知的所有権機関(WIPO)
国際出願番号 2018JP019838
国際公開番号 WO 2018221349
国際出願日 平成30年5月23日(2018.5.23)
国際公開日 平成30年12月6日(2018.12.6)
優先権データ
  • 特願2017-105889 (2017.5.29) JP
発明の名称 (英語) EVALUATION DEVICE, EVALUATION METHOD, PROGRAM, AND INFORMATION RECORDING MEDIUM
発明の概要(英語) The purpose of the present invention is to estimate a score probability from a position time series of players and an object, the players belonging to two teams participating in team sport where a score is obtained by controlling the object until scoring conditions are satisfied. In an evaluation device (101), a distance unit (102) calculates a plurality of kinds of distance time series between an attacker belonging to a desired team, of the two teams, and a defender belonging to the other team on the basis of an attacking segment formed by a position time series while the object is controlled by the desired team. A decomposition unit (103) decomposes the calculated plurality of kinds of distance time series into feature vectors. A learning unit (104) learns for estimating a score probability for an attack segment on the basis of the feature vector by using whether a score is obtained by the attack segment as teaching data.
従来技術、競合技術の概要(英語) BACKGROUND ART
From the prior art, the goal point can be obtained by the waveform of the ball-ball like ball game population, the object score until a condition is satisfied can be controlled by the point can be obtained in a sports team, the player can be analyzed the formation of which has been widely used.
Such as a sports team, a basketball ball object is adopted, soccer, rugby, American football, hockey, or the like or handball, ice hockey and the like employed as a pack the object.
In particular a professional sports industry, the team and the opponent team to analyze the movement of, and evaluation of a strategy or plan, are carried out systematically.
Measured by the development of technology, video camera or from the course of the game each player to collect a time series of the position of the object, various types of information as a series of processes of the distribution has been a service of the system. Of each team coach may, using this system, in accordance with its own experience, such as planning or analysis of the formation can be achieved.
Furthermore, the natural language processing or the like used in the machine learning technique is used, the rough motion of the entire team has been proposed a technique for analyzing the (non-patent document 1, 2).
In non-patent document 1, the trajectory of the players of the neural network (RNN) the optical image input to the recursive type, attack technique for performing the classification of the formation have been disclosed.
In non-patent document 2, the length of the trajectory data of a personal player as a unit, in consideration of the arrangement thereof is not set may be used to, unsupervised learning approach (LDA) is a Dirichlet distribution method applied potential, the formation of the classification of the attack part.
On the other hand, the inventors of the present invention, the control ball in a basketball team is on the side of the attack and a score point condition is satisfied can be obtained in the assessment of whether, and to analyze the relationship between the position of the player. As a result, at (1) least one player belonging to the team on the attack side of the main closest to the ball and the attacker, the defense side player belonging to the team of the user and closest to the attacker the main defense, (a ball-mark distance) the distance between the main, (2) and the attacker, the defense side player belonging to the team as a main operator 2 is the closest attacker defense, the distance between the (ball-help distance) the offense side, (3) the main player belonging to the team other than the each of the sub-attack by the attacker, one of the defense side player belonging to the team of the user closest to each sub-attack and defense, the distance of the maximum value of 3 (path-mark distance) a distance of one, the to-be to play an important role in was found. Based on these parameters whether or not the point can be obtained when estimating whether or not the correct answer ratio, was about 54% (non-patent document 3).
In addition, the present inventors, observability of the infinite dimensional in terms of a subject in a linear operator by the operator person coop spectrum analysis, non-linear dynamic system behavior of the global mode, dynamic mode (Dynamic Mode Decomposition) proposed decomposition algorithm (Non-Patent Document 4). Based on the present algorithm uses the dynamic mode if the kernel method, a potential having a non-linear dynamics to a small number of time series obtained from the system, explicit knowledge of the governing equation of the system even if it is not, strictly mechanical meaning mode can be defined.
  • 出願人(英語)
  • ※2012年7月以前掲載分については米国以外のすべての指定国
  • INSTITUTE OF PHYSICAL AND CHEMICAL RESEARCH
  • 発明者(英語)
  • FUJII, Keisuke
  • KAWAHARA, Yoshinobu
国際特許分類(IPC)
指定国 National States: AE AG AL AM AO AT AU AZ BA BB BG BH BN BR BW BY BZ CA CH CL CN CO CR CU CZ DE DJ DK DM DO DZ EC EE EG ES FI GB GD GE GH GM GT HN HR HU ID IL IN IR IS JO KE KG KH KN KP KR KW KZ LA LC LK LR LS LU LY MA MD ME MG MK MN MW MX MY MZ NA NG NI NO NZ OM PA PE PG PH PL PT QA RO RS RU RW SA SC SD SE SG SK SL SM ST SV SY TH TJ TM TN TR TT TZ UA UG US UZ VC VN ZA ZM ZW
ARIPO: BW GH GM KE LR LS MW MZ NA RW SD SL SZ TZ UG ZM ZW
EAPO: AM AZ BY KG KZ RU TJ TM
EPO: AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR
OAPI: BF BJ CF CG CI CM GA GN GQ GW KM ML MR NE SN ST TD TG

PAGE TOP

close
close
close
close
close
close