18244622. GENERATING PHYSIO-REALISTIC AVATARS FOR TRAINING NON-CONTACT MODELS TO RECOVER PHYSIOLOGICAL CHARACTERISTICS simplified abstract (Microsoft Technology Licensing, LLC)

From WikiPatents
Jump to navigation Jump to search

GENERATING PHYSIO-REALISTIC AVATARS FOR TRAINING NON-CONTACT MODELS TO RECOVER PHYSIOLOGICAL CHARACTERISTICS

Organization Name

Microsoft Technology Licensing, LLC

Inventor(s)

Daniel J. Mcduff of Cambridge MA (US)

Javier Hernandez Rivera of Cambridge MA (US)

Tadas Baltrusaitis of Cambridge (GB)

Erroll William Wood of Cambridge (GB)

GENERATING PHYSIO-REALISTIC AVATARS FOR TRAINING NON-CONTACT MODELS TO RECOVER PHYSIOLOGICAL CHARACTERISTICS - A simplified explanation of the abstract

This abstract first appeared for US patent application 18244622 titled 'GENERATING PHYSIO-REALISTIC AVATARS FOR TRAINING NON-CONTACT MODELS TO RECOVER PHYSIOLOGICAL CHARACTERISTICS

Simplified Explanation

The patent application describes a system and method for generating video sequences with physio-realistic avatars. Here are the key points:

  • The system receives an albedo (surface color) for an avatar.
  • It modifies the sub-surface skin color of the avatar based on physiological data related to a physiologic characteristic.
  • The system then renders an avatar using the modified sub-surface skin color and the albedo.
  • The rendered avatar can be synthesized into a video frame.
  • The video, including the synthesized avatar, can be used to train a machine learning model to detect the physiological characteristic.
  • The machine learning model receives multiple video segments, some of which include synthetic physio-realistic avatars with the physiological characteristic.
  • The model is trained using these video segments.
  • The trained model can be provided to a requesting entity.

Potential applications of this technology:

  • Virtual reality and augmented reality applications can benefit from realistic avatars that accurately represent physiological characteristics.
  • Medical training simulations can use physio-realistic avatars to simulate patient conditions and train healthcare professionals.
  • Gaming and entertainment industries can create more immersive experiences with realistic avatars.
  • Facial recognition systems can be improved by training them with video segments containing avatars that mimic physiological characteristics.

Problems solved by this technology:

  • Traditional methods of generating avatars may not accurately represent physiological characteristics, limiting their realism and applicability.
  • Training machine learning models to detect physiological characteristics often requires real-world data, which may be limited or difficult to obtain.
  • This technology solves these problems by generating synthetic physio-realistic avatars and using them to train machine learning models.

Benefits of this technology:

  • The system allows for the creation of video sequences with highly realistic avatars that accurately represent physiological characteristics.
  • By using synthetic avatars, the system can generate a wide range of physiologic characteristics without relying on real-world data.
  • The trained machine learning models can be used for various applications, such as healthcare, entertainment, and security, improving their accuracy and performance.


Original Abstract Submitted

Systems and methods are provided that are directed to generating video sequences including physio-realistic avatars. In examples, an albedo for an avatar is received, a sub-surface skin color associated with the albedo is modified based on physiological data associated with physiologic characteristic, and an avatar based on the albedo and the modified sub-surface skin color is rendered. The rendered avatar may then be synthesized in a frame of video. In some examples, a video including the synthesized avatar may be used to train a machine learning model to detect a physiological characteristic. The machine learning model may receive a plurality of video segments, where one or more of the video segments includes a synthetic physio-realistic avatar generated with the physiological characteristic. The machine learning model may be trained using the plurality of video segments. The trained model may be provided to a requesting entity.

Retrieved from "http://wikipatents.org/index.php?title=18244622._GENERATING_PHYSIO-REALISTIC_AVATARS_FOR_TRAINING_NON-CONTACT_MODELS_TO_RECOVER_PHYSIOLOGICAL_CHARACTERISTICS_simplified_abstract_(Microsoft_Technology_Licensing,_LLC)&oldid=17104"