Receipt Printer Data Analysis via ML Classification

Solution Overview

Problem

Existing receipt printer analysis methods are inefficient when the layout of printed data is unknown, as they require applying all possible analysis methods, leading to prolonged processing times to determine the appropriate result.

Innovation Solution

An information processing apparatus that uses a machine-learned learning model to determine the type of printed matter based on text data, prioritizing script selection and application to efficiently process print data, including scripts that associate item names with parameters, thereby reducing processing time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If analysis is performed with all possible methods to handle unknown layout, then analysis completeness is improved, but processing time increases significantly

Engineering Contradiction:

Improveanalysis completenessVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary classification of printed matter types using machine learning before applying analysis methods. By determining the type of printed matter in advance (e.g., invoice, receipt, statement), the system can select the appropriate analysis method beforehand, avoiding the need to try all possible methods and significantly reducing processing time while maintaining analysis completeness.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system changes the approach parameter based on the determined printed matter type. Instead of using a fixed set of analysis methods for all cases, the system adjusts which analysis methods are applied according to the identified type, enabling efficient processing tailored to each specific printed matter category.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If all possible analysis methods are applied to ensure comprehensive coverage, then analysis thoroughness is improved, but system efficiency deteriorates

Engineering Contradiction:

Improveanalysis thoroughnessVSAvoidsystem efficiency

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The system performs preliminary classification of printed matter types using machine learning before applying analysis methods. By determining the type of printed matter in advance (e.g., invoice, receipt, statement), the system can select the appropriate analysis method beforehand, avoiding the need to try all possible methods and significantly reducing processing time while maintaining analysis completeness.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system applies different analysis methods tailored to specific printed matter types rather than using a uniform approach for all cases. Each type of printed matter receives the analysis treatment most suitable for its characteristics, improving both thoroughness and efficiency by avoiding unnecessary analysis methods on inappropriate data types.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS10922931B2Information processing apparatus, receipt printer, and information processing method

Publication Date: 2021.02.16 SEIKO EPSON CORP

AI summary

An information processing apparatus is configured which includes an acquisition unit that acquires print data from a receipt printer that prints a printed matter based on the print data, a determination unit that determines a type of the printed matter based on text data included in the print data, and an application unit that selects a script corresponding to the type from a plurality of scripts and applies the selected script to the text data.