公开(公告)号：US20250095406A1

公开(公告)日：2025-03-20

申请号：US18970431

申请日：2024-12-05

Applicant: Google LLC

Inventor： Cem Kemal Hamami ,  Philip Andrew Mansfield ,  Samuel Paradis ,  Michael Williams ,  Wen-Sheng Chu

IPC: G06V40/50 ,  G06V10/762 ,  G06V40/16

Abstract: This document describes systems and techniques that enable continuous personalization of face authentication. In aspects, an authentication system associated with a network includes an authentication manager. The authentication manager receives an embedding representing image data associated with a user's face. The authentication manager generates a confidence score based on the embedding. Further, the authentication manager updates previously enrolled embeddings with the embedding based on the confidence score, the embedding meeting a clustering confidence threshold. Through such a technique, the authentication manager can alter the previously enrolled embeddings by which a future embedding is used to authenticate the user's face. By so doing, the techniques may provide more-accurate and successful user authentication over time.

公开(公告)号：US12189738B2

公开(公告)日：2025-01-07

申请号：US17439770

申请日：2019-09-09

Applicant: Google LLC

Inventor： Michael Moreno ,  Michael Williams ,  Ji Soo Shin ,  Madhi Hamzeh

IPC: H04L29/06 ,  G06F8/654 ,  G06F9/445 ,  G06F21/32 ,  G06F21/57 ,  G06N20/20 ,  G06F16/21

Abstract: This document describes techniques and systems that enable face authentication embedding migration and drift-compensation. The techniques and systems include a user device that is updated to include both a current version of firmware and an updated version of the firmware. Then, an indication of a face-authentication attempt is received along with image data associated with a user's face. After successful authentication, using the current version of firmware on the image data, the user device uses the updated version of the firmware on the same image data to generate a new embedding. The new embedding is stored as part of a migration profile for the user. Additional new embeddings are collected over a series of subsequent face-authentication attempts until a complete set of new embeddings is stored for the migration profile. Then, the old profile is deleted and the migration profile becomes the enrollment profile used for face authentication.

公开(公告)号：US20220100831A1

公开(公告)日：2022-03-31

申请号：US17439770

申请日：2019-09-09

Applicant: Google LLC

Inventor： Michael Moreno ,  Michael Williams ,  Ji Soo Shin ,  Madhi Hamzeh

IPC: G06F21/32 ,  G06F8/654 ,  G06F21/57 ,  G06F9/445 ,  G06N20/20

Abstract: This document describes techniques and systems that enable face authentication embedding migration and drift-compensation. The techniques and systems include a user device that is updated to include both a current version of firmware and an updated version of the firmware. Then, an indication of a face-authentication attempt is received along with image data associated with a user's face. After successful authentication, using the current version of firmware on the image data, the user device uses the updated version of the firmware on the same image data to generate a new embedding. The new embedding is stored as part of a migration profile for the user. Additional new embeddings are collected over a series of subsequent face-authentication attempts until a complete set of new embeddings is stored for the migration profile. Then, the old profile is deleted and the migration profile becomes the enrollment profile used for face authentication.