发明公开
EP3222207A1 SYSTEM AND METHOD FOR NON-INTRUSIVE DETECTION AND MONITORING OF PARKINSON'S DISEASE SYMPTOMS
审中-公开
用于非侵入性检测和监测帕金森病症状的系统和方法
- 专利标题: SYSTEM AND METHOD FOR NON-INTRUSIVE DETECTION AND MONITORING OF PARKINSON'S DISEASE SYMPTOMS
- 专利标题(中): 用于非侵入性检测和监测帕金森病症状的系统和方法
-
申请号: EP16305323.4申请日: 2016-03-23
-
公开(公告)号: EP3222207A1公开(公告)日: 2017-09-27
- 发明人: LONGET, Aurélien , GUEGAN, Marie , LAMBERT, Anne
- 申请人: Thomson Licensing
- 申请人地址: 1-5, rue Jeanne d'Arc 92130 Issy-les-Moulineaux FR
- 专利权人: Thomson Licensing
- 当前专利权人: Thomson Licensing
- 当前专利权人地址: 1-5, rue Jeanne d'Arc 92130 Issy-les-Moulineaux FR
- 代理机构: Huchet, Anne
- 主分类号: A61B5/00
- IPC分类号: A61B5/00 ; G06F19/00 ; A61B5/11 ; A61B5/16
摘要:
A method and apparatus for creating an anomalous behavior detection log for reporting anomalous behavior of a monitored individual in an environment are described including training an anomaly detection classifier using first time-stamped event data for an initial period of time and raw sensor data, accepting second time-stamped event data of subsequent behavior from sensors in the environment for a subsequent period of time, logging the second time-stamped event data of subsequent behavior of the monitored individual, mining an action sequence based on a selected subset of mined manipulation patterns, calculating an anomaly score of each mined action sequence using the anomaly detection classifier and creating an anomalous behavior detection log based on the anomaly scores.
信息查询