# US 12,602,895 B2 · Machine learning-based document splitting and labeling in an electronic document system

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

- Number: US 12,602,895 B2
- Document ID: US12602895B2
- Status: Granted
- Kind: B2
- Application number: 18/753,644
- Filed: 2024-06-25
- Published: 2024-10-17
- Granted: 2026-04-14
- Earliest claimed: 2021-06-22
- Assignees: DocuSign, Inc.
- Applicants: DocuSign, Inc.
- Inventors: Shrinivas Kiran Kaza; Taiwo Raphael Alabi; Ashwath Mohan; Varsha Sri Raghavan; Kim Cuong Phung
- Web page: https://patentgrep.com/patent/US12602895B2

## CPC

- G06V 10/225: {based on a marking or identifier characterising the area}
- G06N 20/20: Ensemble learning
- G06V 10/774: Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
- G06V 30/416: Extracting the logical structure, e.g. chapters, sections or page numbers; Identifying elements of the document, e.g. authors
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/045: Combinations of networks
- G06N 3/0475: Generative networks
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An electronic document system can allow users to upload a document package containing multiple individual component documents. Each component document includes a subset of a plurality of pages that are included in the document package. The electronic document system identifies a page of each component document by applying a machine learning model to the document package. The electronic document system partitions the document package into the individual component documents based on the identified pages. For each individual component document, the electronic document system identifies a document topic corresponding to the component document by applying another machine learning model. The electronic document system modifies a user interface to display each component document and corresponding document topic.

## Claims

### Claim 1

A method, comprising:
receiving a document package from a document package source, the document package comprising a set of pages for a set of component documents, wherein a component document of the set of component documents is prepared for a digital signature;
identifying a subset of pages from the set of pages for a component document of the set of component documents using a first machine learning model;
identifying a document topic for the component document using a second machine learning model; and
presenting the component document and corresponding document topic on a user interface of an electronic display.

### Claim 2 (depends on 1)

The method of claim 1, comprising:
determining a presence of an unrecognized page from the set of pages of the document package;
determining the unrecognized page is for the component document using a third machine learning model; and
assigning the unrecognized page to the component document.

### Claim 3 (depends on 1)

The method of claim 1, comprising:
determining a presence of an unrecognized page from the set of pages of the document package;
determining the unrecognized page is for the component document based on assignment information; and
assigning the unrecognized page to the component document.

### Claim 4 (depends on 1)

The method of claim 1, comprising retrieving the document identification rules from a package template associated with a category for the document package.

### Claim 5 (depends on 1)

The method of claim 1, comprising performing actions on the document stacks based on a package template associated with a category for the document package.

### Claim 6 (depends on 1)

The method of claim 1, comprising updating the document identification rules based on the assignment of the unrecognized page to the identified document stack.

### Claim 7

A non-transitory computer-readable storage medium storing executable instructions that, when executed by a hardware processor, cause the hardware processor to perform operations comprising:
receive a document package from a document package source, the document package comprising a set of pages for a set of component documents, wherein a component document of the set of component documents is prepared for a digital signature;
identify a subset of pages from the set of pages for a component document of the set of component documents using a first machine learning model;
identify a document topic for the component document using a second machine learning model; and
present the component document and corresponding document topic on a user interface of an electronic display.

### Claim 8 (depends on 7)

The computer-readable storage medium of claim 7, comprising instructions that, when executed by the hardware processor, cause the hardware processor to perform operations comprising:
determine a presence of an unrecognized page from the set of pages of the document package;
determine the unrecognized page is for the component document using a third machine learning model; and
assign the unrecognized page to the component document.

### Claim 9 (depends on 7)

The computer-readable storage medium of claim 7, comprising instructions that, when executed by the hardware processor, cause the hardware processor to perform operations comprising:
determine a presence of an unrecognized page from the set of pages of the document package;
determine the unrecognized page is for the component document based on assignment information; and
assign the unrecognized page to the component document.

### Claim 10 (depends on 7)

The computer-readable storage medium of claim 7, comprising instructions that, when executed by the hardware processor, cause the hardware processor to perform operations comprising retrieve the document identification rules from a package template associated with a category for the document package.

### Claim 11 (depends on 7)

The computer-readable storage medium of claim 7, comprising instructions that, when executed by the hardware processor, cause the hardware processor to perform actions on the document stacks based on a package template associated with a category for the document package.

### Claim 12 (depends on 7)

The computer-readable storage medium of claim 7, comprising instructions that, when executed by the hardware processor, cause the hardware processor to perform operations comprising update the document identification rules based on the assignment of the unrecognized page to the identified document stack.

### Claim 13

An apparatus, comprising:
circuitry; and
memory operably coupled to the circuitry, the memory to store instructions that, when executed by the circuitry, causes the circuitry to:
receive a document package from a document package source, the document package comprising a set of pages for a set of component documents, wherein a component document of the set of component documents is prepared for a digital signature;
identify a subset of pages from the set of pages for a component document of the set of component documents using a first machine learning model;
identify a document topic for the component document using a second machine learning model; and
present the component document and corresponding document topic on a user interface of an electronic display.

### Claim 14 (depends on 13)

The apparatus of claim 13, the circuitry to:
determine a presence of an unrecognized page from the set of pages of the document package;
determine the unrecognized page is for the component document using a third machine learning model; and
assign the unrecognized page to the component document.

### Claim 15 (depends on 13)

The apparatus of claim 13, the circuitry to:
determine a presence of an unrecognized page from the set of pages of the document package;
determine the unrecognized page is for the component document based on assignment information; and
assign the unrecognized page to the component document.

### Claim 16 (depends on 13)

The apparatus of claim 13, the circuitry to retrieve the document identification rules from a package template associated with a category for the document package.

### Claim 17 (depends on 13)

The apparatus of claim 13, the circuitry to perform actions on the document stacks based on a package template associated with a category for the document package.