Receipt Parsing System Using Weighted Data Merging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional business expense tracking and reporting systems face challenges due to the variance in receipt formatting styles and poor quality printing materials, leading to limited success with optical character recognition (OCR) technology and incomplete data, which hinders the generation of accurate and comprehensive expense reports.
Innovation Solution
A computer system is developed to automatically merge expense report information from various data sources, including financial institutions and user inputs, using a merging component that assigns weights to elements based on their types and sources, and incorporates OCR and receipt parsing components to extract and process receipt information, thereby generating reliable expense reports.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If OCR technology is used to extract information from receipt images, then automation is improved, but measurement precision deteriorates due to irregular formatting and poor quality printing
Solution Approach 1:
The patent introduces an intermediary processing layer between the receipt image and the final expense report. This layer includes multiple data sources (OCR component, financial institution interface, partner interface, user input) that all feed into a merging component. The merging component acts as a mediator that reconciles discrepancies between these sources, selecting the most reliable data for each expense element based on pre-assigned weights, thus improving overall accuracy while maintaining automation.
Solution Approach 2:
The system dynamically changes parameters by assigning different weights to different data sources based on their reliability for specific expense elements. The merging component evaluates multiple parameters (data source type, element type, confidence scores) and selects the optimal data source for each element, adapting to the specific characteristics of each receipt and data source rather than using a fixed approach.
2Loss of information
If multiple data sources are integrated to improve data completeness, then information completeness is improved, but device complexity increases
Solution Approach 1:
The patent segments the data integration system into distinct, modular components: OCR component for image processing, financial institution interface for transaction data, partner interface for additional information, and user input interface for manual entry. Each component handles a specific data source independently, and the merging component integrates them through a standardized process of identifying expense elements, assigning weights, and selecting the best data source for each element.
Solution Approach 2:
The merging component serves multiple functions: it receives data from various sources, identifies expense elements in standardized formats, assigns weights to different data sources based on their reliability, resolves conflicts between sources, and generates the final expense report. This multi-functional component handles all data integration tasks through a unified approach, reducing overall system complexity despite the multiple data sources.
3Productivity
If conventional OCR is applied to receipts with variance in formatting styles, then processing speed is improved, but manufacturing precision deteriorates due to formatting irregularities
Solution Approach 1:
The system performs preliminary actions by pre-assigning weights to different data sources based on their expected reliability for specific expense elements before processing begins. The merging component has pre-established criteria for evaluating and selecting data sources, allowing it to quickly resolve conflicts without extensive analysis during the actual processing phase, thus maintaining speed while improving accuracy.
Data Source
AI summary
According to another aspect, a computer system is provided. The computer system includes a memory; at least one processor in data communication with the memory; an optical character recognition (OCR) component executable by the at least one processor; and a receipt parsing component executable by the at least processor. The receipt parsing component is configured to receive an image of a receipt; request execution of the OCR component to convert the image to text; identify a value of a vendor element in the text; identify values of additional elements in the text based on the value of the vendor element; and store the vendor elements and the additional elements in a data store.