Granted Patent B2

US 12,511,405 · App. 18/672,243 · Granted Dec 30, 2025

Inventors: Akash Lokesh (Bangalore, IN); Karthik Venkatesh (Bangalore, IN)

Assignee: Dell Products L.P.

G06F21/60G06V20/52H04L12/1827G06F2221/2125

Abstract

Claims (66)

1 . A method for managing security of a user, the method comprising:

analyzing a dataset to annotate data, wherein the data is annotated to mark a region of a sensitive object (SO) and to obtain annotated data;

cleaning the annotated data to obtain cleaned annotated data, wherein at least training data is generated from the cleaned annotated data;

generating, based on a target parameter, an object detection model by training a model using at least the training data;

converting the object detection model into a microservice, wherein the microservice is deployed to a client used by the user;

after the microservice is deployed to the client:

establishing a connection with a set of hardware peripherals of the client, wherein the set of hardware peripherals comprises at least a visual sensor;

capturing, via the visual sensor, a snapshot of an environment that hosts the client, wherein the snapshot is a viewable area of the environment that can be imaged by the visual sensor;

making, based on the snapshot, a first determination that an image of a second SO (SSO) is detected in the environment, wherein the user is notified about the SSO;

receiving, in response to notifying the user, a preference about the SSO from the user;

protecting, based on the preference, the SSO by blurring or masking the image;

making, after protecting the SSO, a second determination that the user joins a call;

monitoring, based on the second determination, the environment to capture a real-time video feed, wherein the video feed is analyzed to obtain an analyzed video feed (AVF);

making, based on the AVF, a third determination that a sensitive action is detected;

protecting, based on the third determination, a third SO (TSO) resulting from the sensitive action by blurring or masking a second image of the TSO, wherein the user is notified about the sensitive action and the TSO;

making, in response to notifying the user, a fourth determination that the user requests removal of the blurring or the masking of the second image; and

terminating, based on the fourth determination, protection of the TSO.

2 . The method of claim 1 , wherein the environment is a public environment or a private environment.

3 . The method of claim 1 , wherein the SSO is a credit card, wherein at least a part of a third image of the credit card is blurred to hide personal information of the user from a second user in the call.

4 . The method of claim 1 , wherein the SSO is a notepad, wherein at least a part of a third image of the notepad is masked to hide mission critical information from a second user in the call, wherein the third image of the notepad is masked by replacing the mission critical information with fake information.

5 . The method of claim 1 , wherein the sensitive action is a change in a body posture of the user, wherein the user changes the body posture to read a notification from a smart phone.

6 . The method of claim 1 , wherein the preference is a security preference that is determined by the user based on a threat detection and mitigation policy.

7 . The method of claim 1 , wherein the TSO is at least one selected from a group consisting of a credit card, a display of a smart phone, a display of a laptop, a notepad, and a keyboard.

8 . The method of claim 1 , wherein the target parameter specifies detecting images of SOs in a second environment that hosts a second user's client.

9 . The method of claim 1 , wherein the set of hardware peripherals further comprises a microphone, an audio sensor, and an electromagnetic radiation sensor.

10 . A method for managing security of a user, the method comprising:

establishing a connection with a set of hardware peripherals of a client, wherein the set of hardware peripherals comprises at least a visual sensor;

capturing, via the visual sensor, a snapshot of an environment that hosts the client, wherein the snapshot is a viewable area of the environment that can be imaged by the visual sensor;

making, based on the snapshot, a first determination that an image of a sensitive object (SO) is detected in the environment, wherein the user is notified about the SO;

receiving, in response to notifying the user, a preference about the SO from the user;

protecting, based on the preference, the SO by blurring or masking the image;

making, after protecting the SO, a second determination that the user joins a call;

monitoring, based on the second determination, the environment to capture a real-time video feed, wherein the video feed is analyzed to obtain an analyzed video feed (AVF);

making, based on the AVF, a third determination that a sensitive action is detected;

protecting, based on the third determination, a second SO (SSO) resulting from the sensitive action by blurring or masking a second image of the SSO, wherein the user is notified about the sensitive action and the SSO;

making, in response to notifying the user, a fourth determination that the user requests removal of the blurring or the masking of the second image; and

terminating, based on the fourth determination, protection of the SSO.

11 . The method of claim 10 , further comprising:

prior to establishing the connection:

analyzing a dataset to annotate data, wherein the data is annotated to mark a region of a third SO (TSO) and to obtain annotated data;

cleaning the annotated data to obtain cleaned annotated data, wherein at least training data is generated from the cleaned annotated data;

generating, based on a target parameter, an object detection model by training a model using at least the training data; and

converting the object detection model into a microservice, wherein the microservice is deployed to the client used by the user.

12 . The method of claim 11 , wherein the TSO is at least one selected from a group consisting of a credit card, a display of a smart phone, a display of a laptop, a notepad, and a keyboard.

13 . The method of claim 11 , wherein the target parameter specifies detecting images of SOs in a second environment that hosts a second user's client.

14 . The method of claim 10 , wherein the sensitive action is a change in a body posture of the user, wherein the user changes the body posture to read a notification from a smart phone.

15 . The method of claim 10 , wherein the set of hardware peripherals further comprises a microphone, an audio sensor, and an electromagnetic radiation sensor.

16 . The method of claim 10 , wherein the preference is a security preference that is determined by the user based on a threat detection and mitigation policy.

17 . The method of claim 10 , wherein the SO is a credit card, wherein at least a part of a third image of the credit card is blurred to hide personal information of the user from a second user in the call.

18 . The method of claim 10 , wherein the SO is a notepad, wherein at least a part of a third image of the notepad is masked to hide mission critical information from a second user in the call, wherein the third image of the notepad is masked by replacing the mission critical information with fake information.

19 . A method for managing security of a user, the method comprising:

establishing a connection with a set of hardware peripherals of a client, wherein the set of hardware peripherals comprises at least a visual sensor;

capturing, via the visual sensor, a snapshot of an environment that hosts the client, wherein the snapshot is a viewable area of the environment that can be imaged by the visual sensor;

making, based on the snapshot, a first determination that an image of a sensitive object (SO) is detected in the environment, wherein the user is notified about the SO;

receiving, in response to notifying the user, a preference about the SO from the user, wherein the reference specifies taking no action with respect to the image;

making, after the receiving, a second determination that the user joined a call;

monitoring, based on the second determination, the environment to capture a real-time video feed, wherein the video feed is analyzed to obtain an analyzed video feed (AVF);

making, based on the AVF, a third determination that a sensitive action is detected;

protecting, based on the third determination, a second SO (SSO) resulting from the sensitive action by blurring or masking a second image of the SSO, wherein the user is notified about the sensitive action and the SSO; and

making, in response to notifying the user, a fourth determination that the user does not request removal of the blurring or the masking of the second image.

20 . The method of claim 19 , further comprising:

prior to establishing the connection:

analyzing a dataset to annotate data, wherein the data is annotated to mark a region of a third SO (TSO) and to obtain annotated data;

cleaning the annotated data to obtain cleaned annotated data, wherein at least training data is generated from the cleaned annotated data;

generating, based on a target parameter, an object detection model by training a model using at least the training data; and

converting the object detection model into a microservice, wherein the microservice is deployed to the client used by the user.

Assignments (1)

Continuity (1)

References Cited (24)