Patent · US Active

Systems and methods for sensor data analysis through machine learning

US9818126B1 · kind B1 · utility

7Cited by
0References
24Claims
0Family size

Assignee

Inventors

Key dates

Filing dateApr 20, 2016
Grant dateNov 14, 2017
Priority date
Expiry dateApr 20, 2036

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06V40/18
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

Sensor data analysis may include obtaining video data, detecting facial data within the video data, extracting the facial data from the video data, detecting indicator data within the video data, extracting the indicator data from the video data, transforming the extracted facial data into representative facial data, and determining a mood of the person by associating learned mood indicators derived from other detected facial data with the representative facial data. The analysis may include determining that the representative facial data is associated with a complex profile, and determining a context regarding the person within the environment by weighting and processing the determined mood, at least one subset of data representing information about the person of the complex profile, and the indicator data. The analysis may include determining a user experience for the person, and communicating the determined user experience to a device associated with the person.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.