# US 11,443,541 B2 · Classification of person type in a visual medium

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 11,443,541 B2
- Document ID: US11443541B2
- Status: Granted
- Kind: B2
- Application number: 17/006,468
- Filed: 2020-08-28
- Published: 2022-03-03
- Granted: 2022-09-13
- Earliest claimed: 2020-08-28
- Assignees: Sensormatic Electronics Corporation
- Applicants: Sensormatic Electronics, LLC
- Inventors: Michael C. Stewart; Joseph Celi; Gopi Subramanian
- Web page: https://patentgrep.com/patent/US11443541B2

## CPC

- G06V 40/103: {Static body considered as a whole, e.g. static pedestrian or occupant recognition}
- G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06F 18/217: {Validation; Performance evaluation; Active pattern learning techniques}
- G06F 18/2433: Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G06T 7/11: Region-based segmentation
- G06V 10/764: using classification, e.g. of video objects
- G06V 20/52: Surveillance or monitoring of activities, e.g. for recognising suspicious objects (recognising microscopic objects G06V20/69)
- G06T 2207/20081: Training; Learning
- G06V 2201/09: Recognition of logos
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

Disclosed herein are apparatuses and methods for classifying person type in a visual medium. The implementations include identifying a person in an image, generating a visual boundary around the identified person, and dividing the image into a plurality of image portions such that multiple images portions partially depict the identified person inside the visual boundary. The implementations include generating the divided image on a user interface that prompts a user to select each portion that at least partially depicts a person of a predetermined type. The implementations include receiving a selection of at least one portion. The implementations include generating, for a machine learning model and in response to determining that the selection comprises the identified person, a training input that comprises a training image and a tag indicating that the identified person is of the predetermined type. Additionally, the implementations include storing the training input in memory.

## Claims

### Claim 1

A method for classifying person type in a visual medium, comprising:
identifying a person in an image using image classification;
generating, on the image, a visual boundary around the identified person;
dividing the image into a plurality of image portions such that multiple images portions partially depict the identified person inside the visual boundary;
generating, for display, the plurality of image portions on a user interface that prompts a user to select each portion, of the plurality of image portions, that at least partially depicts a person of a predetermined type;
receiving a selection of at least one portion in the divided image via the user interface;
generating, for a machine learning model and in response to determining that the selection comprises the identified person inside the visual boundary, a training input that comprises a training image depicting contents inside the visual boundary and a tag indicating that the identified person inside the visual boundary is of the predetermined type; and
storing the training input in memory.

### Claim 2 (depends on 1)

The method of claim 1, wherein the machine learning model is a one-class support vector machine.

### Claim 3 (depends on 1)

The method of claim 1, further comprising:
vectorising the training input to create a training vector;
training the machine learning model to identify the person of the predetermined type in an input image using the training vector.

### Claim 4 (depends on 3)

The method of claim 3, further comprising:
identifying another person in a different image using image classification; and
determining, using the machine learning model, whether the another person is of the predetermined type.

### Claim 5 (depends on 4)

The method of claim 4, wherein the different image is captured by a camera, further comprising:
monitoring, in response to determining that the another person is of the predetermined type, the another person using the camera until the another person exists a field of view of the camera;
generating statistics associated with the another person based on the monitoring; and
transmitting the statistics to a device.

### Claim 6 (depends on 4)

The method of claim 4, further comprising:
receiving an output from the machine learning model indicating that the another person is of the predetermined type; and
verifying the output by:
dividing the different image into a plurality of different image portions;
generating, for display, the plurality of different image portions on the user interface;
receiving another selection of at least one portion of the plurality of divided different image portions via the user interface, wherein the another selection is indicative of the another person being of the predetermined type; and
generating, in response to determining that the another selection comprises the another person, another training input that comprises another training image depicting the another person and another tag indicating that the another person is of the predetermined type.

### Claim 7 (depends on 6)

The method of claim 6, wherein the machine learning model further outputs a confidence score indicative of a probability that the output is correct, and wherein the verifying is performed in response to determining that the confidence score is below a threshold confidence score.

### Claim 8 (depends on 3)

The method of claim 3, wherein the predetermined type is an employee, and wherein vectorising the stored training input to create the training vector further comprises:
identifying at least one feature of the training input for inclusion in the training vector, wherein the at least one feature is one or more of:
an outline of an employee uniform,
a color representation of the employee uniform,
a logo on the employee uniform, or
presence of a nametag.

### Claim 9 (depends on 1)

The method of claim 1, further comprising:
determining, prior to dividing the image, whether at least a predetermined amount of a body of the identified person is depicted in the image;
discarding the image in response to determining that at least the predetermined amount of the body is not depicted; and
identifying the person in a different image that depicts at least the predetermined amount of the body of the identified person.

### Claim 10

An apparatus for classifying person type in a visual medium, comprising:
a memory; and
a processor in communication with the memory and configured to:
identify a person in an image using image classification;
generate, on the image, a visual boundary around the identified person;
divide the image into a plurality of image portions such that multiple images portions partially depict the identified person inside the visual boundary;
generate, for display, the plurality of image portions on a user interface that prompts a user to select each portion, of the plurality of image portions, that at least partially depicts a person of a predetermined type;
receive a selection of at least one portion in the divided image via the user interface;
generate, for a machine learning model and in response to determine that the selection comprises the identified person inside the visual boundary, a training input that comprises a training image depicting contents inside the visual boundary and a tag indicating that the identified person inside the visual boundary is of the predetermined type; and
store the training input in memory.

### Claim 11 (depends on 10)

The apparatus of claim 10, wherein the machine learning model is a one-class support vector machine.

### Claim 12 (depends on 10)

The apparatus of claim 10, wherein the processor is further configured to:
vectorising the training input to create a training vector;
train the machine learning model to identify the person of the predetermined type in an input image using the training vector.

### Claim 13 (depends on 12)

The apparatus of claim 12, wherein the processor is further configured to:
identify another person in a different image using image classification; and
determine, using the machine learning model, whether the another person is of the predetermined type.

### Claim 14 (depends on 13)

The apparatus of claim 13, wherein the different image is captured by a camera, wherein the processor is further configured to:
monitor, in response to determine that the another person is of the predetermined type, the another person using the camera until the another person exists a field of view of the camera;
generate statistics associated with the another person based on the monitoring; and
transmit the statistics to a device.

### Claim 15 (depends on 13)

The apparatus of claim 13, wherein the processor is further configured to:
receive an output from the machine learning model indicating that the another person is of the predetermined type; and
verify the output by:
dividing the different image into a plurality of different image portions;
generating, for display, the plurality of different image portions on the user interface;
receiving another selection of at least one portion of the plurality of divided different image portions via the user interface, wherein the another selection is indicative of the another person being of the predetermined type; and
generating, in response to determine that the another selection comprises the another person, another training input that comprises another training image depicting the another person and another tag indicating that the another person is of the predetermined type.

### Claim 16 (depends on 15)

The apparatus of claim 15, wherein the machine learning model further outputs a confidence score indicative of a probability that the output is correct, and wherein the verifying is performed in response to determine that the confidence score is below a threshold confidence score.

### Claim 17 (depends on 12)

The apparatus of claim 12, wherein the predetermined type is an employee, and wherein vectorising the stored training input to create the training vector the processor is further configured to:
identify at least one feature of the training input for inclusion in the training vector, wherein the at least one feature is one or more of:
an outline of an employee uniform,
a color representation of the employee uniform,
a logo on the employee uniform, or
presence of a nametag.

### Claim 18 (depends on 10)

The apparatus of claim 10, wherein the processor is further configured to:
determine, prior to divide the image, whether at least a predetermined amount of a body of the identified person is depicted in the image;
discard the image in response to determine that at least the predetermined amount of the body is not depicted; and
identify the person in a different image that depicts at least the predetermined amount of the body of the identified person.

### Claim 19

A non-transitory computer readable medium for classifying person type in a visual medium, executable by a processor to:
identify a person in an image using image classification;
generate, on the image, a visual boundary around the identified person;
divide the image into a plurality of image portions such that multiple images portions partially depict the identified person inside the visual boundary;
generate, for display, the plurality of image portions on a user interface that prompts a user to select each portion, of the plurality of image portions, that at least partially depicts a person of a predetermined type;
receive a selection of at least one portion in the divided image via the user interface;
generate, for a machine learning model and in response to determine that the selection comprises the identified person inside the visual boundary, a training input that comprises a training image depicting contents inside the visual boundary and a tag indicating that the identified person inside the visual boundary is of the predetermined type; and
store the training input in memory.