# US 12,182,757 B2 · Automated inspection of closed package contents and scanning to optically coupled systems

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

- Number: US 12,182,757 B2
- Document ID: US12182757B2
- Status: Granted
- Kind: B2
- Application number: 18/478,695
- Filed: 2023-09-29
- Published: 2024-01-25
- Granted: 2024-12-31
- Earliest claimed: 2021-12-28
- Assignees: Vicente Cardenas
- Applicants: Vicente Cardenas
- Inventors: Vicente Cardenas; Raul Trevino
- Web page: https://patentgrep.com/patent/US12182757B2

## CPC

- G06Q 10/00: Administration; Management
- G06Q 10/0833: Tracking
- G06Q 10/08: Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q 10/0832: Special goods or special handling procedures, e.g. handling of hazardous or fragile goods
- G06Q 10/0838: {Historical data}
- G06Q 10/0841: {Compliance with shipping requirements}
- G06T 7/0002: {Inspection of images, e.g. flaw detection}
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

Apparatus and associated methods relate to automatically generating a jurisdiction entry compliance indicator (JECI) based on an automatically generated content attributes profile (CAP) for a closed package. In an illustrative example, the CAP may be generated based on predetermined content attributes selected by a content characterization model as a function of a contents description profile (CDP) for the package. The CDP may, for example, be automatically generated by a natural language processing model as a function of a predetermined text description of contents of the package. The JECI may be generated as a function of the CAP, predetermined permissions rules, and predetermined permissions attributes. The predetermined permissions attributes may be associated with the originator. The predetermined permissions rules may be identified by a rules identification model as a function of the CAP and a destination of the package. Various embodiments may advantageously automatically inspect contents of closed packages.

## Claims

### Claim 1

A computer program product comprising:
a program of instructions tangibly embodied on a non-transitory computer readable medium wherein, when the instructions are executed on a processor, the processor causes operations to be performed to automatically inspect contents of closed packages, the operations comprising:
apply, in response to receiving a signal from a customs broker package inspection station corresponding to text that indicates contents of a closed package, the text being external to the package, a natural language processing model to generate a content description profile, wherein the signal further indicates an originator of the package and a destination jurisdiction;
perform model training operations comprising;
train a content characterization model with training data to generate a trained content characterization model, the training data comprising predetermined content description data and content attribute data associated with the predetermined content description data;
apply the trained content characterization model to testing data to generate a training metric as a function of a comparison between predicted content attributes and model content attributes associated with the testing data; and,
when the training score does not meet a predetermined criterion, then repeat at least the steps of train and apply;
apply the content characterization model to the content description profile to generate a content attributes profile associating the contents of the package with at least one predetermined content attribute; and,
generate a jurisdiction entry compliance indicator (JECI) as a function of the destination jurisdiction, the content attributes profile, permission attributes, and permission rules such that the JECI indicates a permissibility of the contents of the package to enter the destination jurisdiction;
when the JECI meets at least one predetermined confidence criterion, then automatically generate and transmit an indication to customs broker personnel operating the inspection station not to open the package and perform manual inspection of the contents,
wherein the at least one predetermined confidence criterion corresponds to accuracy of the content attribute characterizing the contents of the closed package and to conformance of the contents of the closed package to the predetermined permissions rules and the predetermined permissions attributes.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, wherein the operations further comprise:
select the predetermined permissions rules and assign the predetermined permission attributes comprising:
determining that the package qualifies for a simplified entry process.

### Claim 3 (depends on 1)

The computer program product of claim 1, further comprising:
determine, based on the signal corresponding to text, whether the package corresponds to a simplified inspection process; and,
when the package does not correspond to the simplified inspection process, then assign at least one of the predetermined permission attributes to define a permit required by the destination jurisdiction as a function of the at least one predetermined content attribute.

### Claim 4 (depends on 3)

The computer program product of claim 3, wherein:
the JECI is determined to not meet the at least one predetermined confidence criterion if any permit required is not associated with at least one corresponding predetermined permit record stored in a database in association with the originator of the package.

### Claim 5 (depends on 1)

The computer program product of claim 1, further comprising:
when the JECI is between the at least one predetermined confidence criterion and a second predetermined confidence criterion, then generate an indication to manually inspect the package.

### Claim 6 (depends on 1)

The computer program product of claim 1, further comprising:
when the JECI is below a second predetermined confidence criterion, then generate an indication to reject the package without manual inspection of the contents.

### Claim 7 (depends on 1)

The computer program product of claim 1, further comprising:
generate a human-readable display based on the indication to pass the package, wherein the human-readable display comprises a predetermined visual indicia prompting a user to pass the package.

### Claim 8 (depends on 1)

The computer program product of claim 1, further comprising perform model training operations, the model training operations comprising:
receive an initial content characterization model;
retrieve training data comprising historic content description data and historic content attribute data associated with the historic content description data;
divide the training data into a training set and a test set;
train the initial content characterization model with the training data to generate a trained content characterization model;
apply the trained content characterization model to the historic content description data of the test set to generate predicted content attributes;
compare the predicted content attributes to the historic content attribute data of the test set; and,
when the predicted content attributes are not within a predetermined matching criterion, then receive additional training data and repeat at least the steps of train, apply, and compare.

### Claim 9

A computer-implemented method performed by at least one processor to automatically inspect contents of closed packages, the method comprising:
apply, in response to receiving a signal from a customs broker package inspection station corresponding to text that indicates contents of a closed package, the text being external to the package, a natural language processing model to generate a content description profile, wherein the signal further indicates an originator of the package and a destination jurisdiction;
perform model training operations comprising;
train an initial content characterization model with training data to generate a trained content characterization model, the training data comprising predetermined content description data and content attribute data associated with the predetermined content description data;
apply the trained content characterization model to testing data to generate a training metric as a function of a comparison between predicted content attributes and model content attributes associated with the testing data; and,
when the training score does not meet a predetermined criterion, then repeat at least the steps of train and apply;
apply the content characterization model to the content description profile to generate a content attributes profile associating the contents of the package with at least one predetermined content attribute; and,
generate a jurisdiction entry compliance indicator (JECI) as a function of the destination jurisdiction, the content attributes profile, permission attributes, and permission rules such that the JECI indicates a permissibility of the contents of the package to enter the destination jurisdiction;
when the JECI meets at least one predetermined confidence criterion, then automatically generate and transmit an indication to a customs broker personnel operating the inspection station not to open the package and perform-manual inspection of the contents,
wherein the at least one predetermined confidence criterion corresponds to an accuracy of the content attribute characterizing the contents of the closed package and a conformance of the contents of the closed package to the predetermined permissions rules and the predetermined permissions attributes.

### Claim 10 (depends on 9)

The computer-implemented method of claim 9, wherein the operations further comprise:
select the predetermined permissions rules and assign the predetermined permission attributes comprising:
determining that the package qualifies for a simplified entry process.

### Claim 11 (depends on 9)

The computer-implemented method of claim 9, further comprising:
determine, based on the signal corresponding to text, whether the package corresponds to a simplified inspection process; and,
when the package does not correspond to the simplified inspection process, then assign at least one of the predetermined permission attributes to define a permit required by the destination jurisdiction as a function of the at least one predetermined content attribute.

### Claim 12 (depends on 11)

The computer-implemented method of claim 11, wherein:
the JECI is determined to not meet the at least one predetermined confidence criterion if any permit required is not associated with at least one corresponding predetermined permit record stored in a database in association with the originator of the package.

### Claim 13 (depends on 9)

The computer-implemented method of claim 9, wherein the content characterization model is dynamically trained based on historical content attributes associated with packages corresponding to a JECI that did not meet the at least one predetermined confidence criterion.

### Claim 14 (depends on 9)

The computer-implemented method of claim 9, further comprising:
when the JECI is between the at least one predetermined confidence criterion and a second predetermined confidence criterion, then generate an indication to manually inspect the package.

### Claim 15 (depends on 9)

The computer-implemented method of claim 9, further comprising:
when the JECI is below a second predetermined confidence criterion, then generate an indication to reject the package without manual inspection of the contents.

### Claim 16 (depends on 9)

The computer-implemented method of claim 9, wherein the model training operations comprise:
receive an initial content characterization model;
retrieve the training data comprising historic content description data and historic content attribute data associated with the historic content description data;
divide the training data into a training set and a test set;
train the initial content characterization model with the training data to generate a trained content characterization model;
apply the trained content characterization model to the historic content description data of the test set to generate predicted content attributes;
compare the predicted content attributes to the historic content attribute data of the test set; and,
when the predicted content attributes are not within a predetermined matching criterion, then receive additional training data and repeat at least the steps of train, apply, and compare,
wherein the predetermined training criterion comprises the predetermined matching criterion.

### Claim 17 (depends on 9)

The computer-implemented method of claim 9, further comprising:
generate a human-readable display based on the indication to pass the package, wherein the human-readable display comprises a predetermined visual indicia prompting a user to pass the package.

### Claim 18 (depends on 17)

The computer-implemented method of claim 17, wherein the predetermined visual indicia comprises a region of a predetermined color corresponding to the indication to pass the package.

### Claim 19

A system comprising:
a non-transitory data store comprising a program of instructions; and,
a processor operably coupled to the data store such that, when the processor executes the program of instructions, the processor causes operations to be performed to automatically inspect contents of closed packages, the operations comprising:
apply, in response to receiving a signal from a customs broker package inspection station corresponding to text external to the package that indicates contents of a closed package, a natural language processing model to generate a content description profile, wherein the signal further indicates an originator of the package and a destination jurisdiction;
perform model training operations comprising;
train an initial content characterization model with training data to generate a trained content characterization model, the training data comprising predetermined content description data and content attribute data associated with the predetermined content description data;
apply the trained content characterization model to testing data to generate a training metric as a function of a comparison between predicted content attributes and model content attributes associated with the testing data; and,
when the training score does not meet a predetermined criterion, then repeat at least the steps of train and apply;
apply the content characterization model to the content description profile to generate a content attributes profile associating the contents of the package with at least one predetermined content attribute; and,
generate a jurisdiction entry compliance indicator (JECI) as a function of the destination jurisdiction, the content attributes profile, permission attributes, and permission rules such that the JECI indicates a permissibility of the contents of the package to enter the destination jurisdiction;
when the JECI meets at least one predetermined confidence criterion, then automatically generate and transmit an indication to a customs broker personnel operating the inspection station not to open the package and perform manual inspection of the contents;
wherein the at least one predetermined confidence criterion corresponds to an accuracy of the content attribute characterizing the contents of the closed package and a conformance of the contents of the closed package to the predetermined permissions rules and the predetermined permissions attributes.

### Claim 20 (depends on 19)

The system of claim 19, wherein the content characterization model is dynamically trained based on historical content attributes associated with packages corresponding to a JECI that did not meet the at least one predetermined confidence criterion.

### Claim 21 (depends on 1)

The computer program product of claim 1, further comprising a second set of operations comprising:
select predetermined permissions rules as a function of the destination jurisdiction and the content attributes profile;
assign predetermined permission attributes as a function of the originator of the package and the content attributes profile;
apply the predetermined permissions rules to the predetermined permission attributes to generate such that the JECI that indicates a permissibility of the contents of the package to enter the destination jurisdiction; and
apply at least one predetermined confidence criterion to the JECI.

### Claim 22 (depends on 1)

The Computer program product of claim 1, further comprising:
dynamically training the content characterization model based on historical content attributes associated with packages corresponding to a JECI that did not meet the at least one predetermined confidence criterion, the training comprising:
train an initial content characterization model with training data to generate a trained content characterization model, the training data comprising predetermined content description data and content attribute data associated with the predetermined content description data;
apply the trained content characterization model to testing data to generate a training metric as a function of a comparison between predicted content attributes and model content attributes associated with the testing data; and,
when the training metric does not meet a predetermined training criterion, then repeat at least the steps of train and apply.

### Claim 23 (depends on 9)

The computer implemented method of claim 9, further comprising a second set of operations comprising:
select predetermined permissions rules as a function of the destination jurisdiction and the content attributes profile;
assign predetermined permission attributes as a function of the originator of the package and the content attributes profile;
apply the predetermined permissions rules to the predetermined permission attributes to generate such that the JECI that indicates a permissibility of the contents of the package to enter the destination jurisdiction; and
apply at least one predetermined confidence criterion to the JECI.

### Claim 24 (depends on 9)

The computer implemented method of claim 9, further including
dynamically training the content characterization model based on historical content attributes associated with packages corresponding to a JECI that did not meet the at least one predetermined confidence criterion, the training comprising:
train an initial content characterization model with training data to generate a trained content characterization model, the training data comprising predetermined content description data and content attribute data associated with the predetermined content description data;
apply the trained content characterization model to testing data to generate a training metric as a function of a comparison between predicted content attributes and model content attributes associated with the testing data; and,
when the training metric does not meet a predetermined training criterion, then repeat at least the steps of train and apply.

### Claim 25 (depends on 19)

The system of claim 19, further comprising a second set of operations comprising:
select predetermined permissions rules as a function of the destination jurisdiction and the content attributes profile;
assign predetermined permission attributes as a function of the originator of the package and the content attributes profile;
apply the predetermined permissions rules to the predetermined permission attributes to generate such that the JECI that indicates a permissibility of the contents of the package to enter the destination jurisdiction; and
apply at least one predetermined confidence criterion to the JECI.

### Claim 26 (depends on 19)

The system of claim 19, further including
dynamically training the content characterization model based on historical content attributes associated with packages corresponding to a JECI that did not meet the at least one predetermined confidence criterion, the training comprising:
train an initial content characterization model with training data to generate a trained content characterization model, the training data comprising predetermined content description data and content attribute data associated with the predetermined content description data;
apply the trained content characterization model to testing data to generate a training metric as a function of a comparison between predicted content attributes and model content attributes associated with the testing data; and,
when the training metric does not meet a predetermined training criterion, then repeat at least the steps of train and apply.