公开(公告)号：US11462218B1

公开(公告)日：2022-10-04

申请号：US16862359

申请日：2020-04-29

Applicant: Amazon Technologies, Inc.

Inventor： Narendra Gyanchandani ,  Bilyana Slavova ,  Njenga Kariuki ,  Matthew Heida ,  Naveenan Vasudevan ,  Stephen Phillip Pant

IPC: G06F1/16 ,  G10L15/22 ,  G10L25/78 ,  G06F1/18 ,  G06F1/3287

Abstract: Described are apparatus, systems, and methods that are operable to periodically detected and record a voice of a user associated with or wearing a wearable device. As described, the wearable device apparatus is configured to transition various components of the device between a low power state and an active state to determine if audio data that includes voice is detected. If voice is not detected, the components are transitioned back to the low power state, thereby conserving power of the wearable device. If voice is detected, the audio data that includes that voice is collected for further processing.

公开(公告)号：US11783834B1

公开(公告)日：2023-10-10

申请号：US17936713

申请日：2022-09-29

Applicant: Amazon Technologies, Inc.

Inventor： Narendra Gyanchandani ,  Bilyana Slavova ,  Njenga Kariuki ,  Matthew Heida ,  Naveenan Vasudevan ,  Stephen Phillip Pant

IPC: G10L15/22 ,  G10L25/78 ,  G06F1/18 ,  G06F1/16 ,  G06F1/3287

CPC classification number: G10L15/22 ,  G06F1/163 ,  G06F1/189 ,  G06F1/3287 ,  G10L25/78

Abstract: Described are apparatus, systems, and methods that are operable to periodically detected and record a voice of a user associated with or wearing a wearable device. As described, the wearable device apparatus is configured to transition various components of the device between a low power state and an active state to determine if audio data that includes voice is detected. If voice is not detected, the components are transitioned back to the low power state, thereby conserving power of the wearable device. If voice is detected, the audio data that includes that voice is collected for further processing.

公开(公告)号：US11632456B1

公开(公告)日：2023-04-18

申请号：US17101769

申请日：2020-11-23

Applicant: Amazon Technologies, Inc.

Inventor： Mark Jared Mevorah ,  Njenga Kariuki

IPC: H04M1/72454 ,  G10L15/22 ,  H04M1/656 ,  G10L15/18 ,  G10L15/04 ,  H04M1/72484

Abstract: Described are systems and methods that detect a call connection of a call (e.g., telephone call, social media call, etc.) between a first device of a user, such as a cellular phone, and a device of another person with which the user is to interact via the call. Upon detection of the call connection, another device of the user (e.g., wearable) records audio of the environment to detect speech output by the user during the call. The speech is then processed to determine sentiment(s) of the user during the call. The determined sentiment(s) of the user may then be presented to the user to aid the user in understanding how the speech of the user is perceived by others.

公开(公告)号：US11430467B1

公开(公告)日：2022-08-30

申请号：US17027509

申请日：2020-09-21

Applicant: Amazon Technologies, Inc.

Inventor： Naveenan Vasudevan ,  Girish Vuggina ,  Njenga Kariuki ,  Siddhi Pathak

IPC: G10L25/30 ,  G10L25/63 ,  G10L15/22

Abstract: Disclosed are systems and methods for determining sentiment data representative of an emotional state of a user during an interaction with one or more non-users. The disclosed implementations may further determine, for frequent interactions between the user and a non-user, the identity of the non-user and provide information to the user relating to the sentiment of the user when interacting with the non-user.