公开(公告)号：US20170140244A1

公开(公告)日：2017-05-18

申请号：US15350962

申请日：2016-11-14

Applicant: Tata Consultancy Services Limited

Inventor： GEETIKA SHARMA ,  GAUTAM SHROFF ,  PUNEET AGARWAL ,  ADITEYA PANDEY ,  GUNJAN SEHGAL ,  KAUSHAL ASHOKBHAI PANERI ,  BRIJENDRA SINGH

IPC: G06K9/62 ,  G06T11/20 ,  G06N99/00

CPC classification number: G06K9/6212 ,  G06F17/30572 ,  G06K9/6221 ,  G06K9/6265 ,  G06N99/005 ,  G06T11/206

Abstract: This disclosure relates generally to multi-sensor visual analytics, and more particularly to method and system for multi-sensor visual analytics using machine-learning models. In one embodiment, a method for multi-sensor visual analytics includes acquiring sensor data associated with a plurality of sensors for a plurality of days of operation. A plurality of multi-dimensional histograms, having operational profiles of the plurality of sensors are computed from the sensor data. The plurality of multi-dimensional histograms are monitored, and a plurality of multi-sensor patterns are obtained from the plurality of multi-dimensional histograms. The plurality of multi-sensor patterns are indicative of one or more properties of a plurality of sensor-clusters of the plurality of sensors. One or more visual analytical tasks are performed by processing the plurality of multi-sensor patterns using at least one machine-learning model. The one or more visual models are rendered based on the processing of the multi-sensor patterns.