Iterative Document Classification via Partial Scanning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional document classification systems scan entire documents, potentially exposing sensitive information and lacking efficiency in classification processes.

Innovation Solution

The system iteratively scans portions of a document, extracts data using OCR, and employs neural networks or machine learning models to determine document types based on confidence thresholds, minimizing exposure of sensitive information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the system scans the entire document to ensure accurate classification, then classification accuracy is improved, but the exposure of sensitive information increases and processing time increases

Engineering Contradiction:

Improveclassification accuracyVSAvoidexposure of sensitive information

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent divides the document into multiple portions and iteratively scans only the necessary portions (e.g., header, footer, or specific sections) rather than the entire document. This segmentation allows the system to classify documents based on critical regions while minimizing exposure of sensitive content in other areas.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent extracts and analyzes only the essential portions of the document that contain classification-relevant information. By taking out and processing only these critical segments (such as document headers, footers, or specific fields), the system achieves accurate classification without unnecessarily exposing or processing sensitive information throughout the entire document.

Inventive Principle:

Principle #2Taking out (Extraction)

2Reliability

If the system scans the entire document to ensure comprehensive analysis, then classification reliability is improved, but processing speed decreases

Engineering Contradiction:

Improveclassification reliabilityVSAvoidprocessing speed

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent performs preliminary scanning of document portions that are most likely to contain classification indicators (such as headers, footers, or metadata sections) before proceeding to full document analysis. This preliminary action allows the system to quickly identify and classify documents that can be reliably classified from partial information, thereby improving processing speed while maintaining reliability for those cases.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent applies partial action by scanning only the necessary portions of documents required for classification rather than the entire document. For many document types, classification can be reliably determined from specific sections alone, making the partial scanning approach both faster and sufficiently reliable for the classification task.

Inventive Principle:

Principle #16Partial or excessive action

3Loss of information

If the system uses iterative scanning with increasing portion size, then the amount of sensitive information exposed is reduced, but the complexity of the classification process increases

Engineering Contradiction:

Improveexposure of sensitive informationVSAvoidcomplexity of classification process

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent implements a dynamic scanning approach where the portion of the document scanned adjusts iteratively based on classification confidence levels. The system starts with a small initial portion and progressively increases the scanned area only when necessary, creating a dynamic and adaptive process that balances sensitivity protection with classification needs while managing complexity through structured progression.

Inventive Principle:

Principle #15Dynamics

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach enhances the speed and accuracy of document classification by reducing the amount of data needed for classification, while protecting sensitive information by scanning only necessary portions of documents.

Implementation Method 1

determine whether the document type is identified by, using a neural network (NN), determining whether the first text data corresponds to one or more document types of a plurality of document types based on one or more confidence thresholds

Methodology Applied

Scientific EffectNeural network processing:

Implementation Method 2

extract first text data from the portion of the first document by performing optical character recognition (OCR) on the portion of the first document

Methodology Applied

Scientific EffectOptical character recognition:

Data Source

PatentUS12306880B2Systems and methods for classifying documents

Publication Date: 2025.05.20 CAPITAL ONE SERVICES LLC

AI summary

A system may iteratively scan a portion of a document, extract first data from the portion of the document, and determine, using a trained model, whether the first data corresponds to one or more document types based on one or more confidence thresholds. The system may repeat this process, increasing the portion of the document scanned by a predetermined amount each iteration, until the first data corresponds to the one or more document types based on the one or more confidence thresholds. Responsive to determining the first data corresponds to the one or more document types based on the one or more confidence thresholds, the system may cause a graphical user interface (GUI) of a user device to display a notification indicating a document type match.