Unsupervised Retail Inventory Audit via Text Recognition

Solution Overview

Problem

Existing systems for automatic product inventory auditing require annotated training data, which is labor-intensive and time-consuming, making them inefficient for generating audit reports in retail settings.

Innovation Solution

A system and method for generating live audit reports using unannotated image data through unsupervised image processing techniques like edge detection and text recognition, implemented in a multi-tenant database system, allowing for real-time product inventory and placement information generation without the need for manual annotation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If annotated training data is used for automatic product inventory auditing, then measurement precision is improved, but loss of time and loss of substance increase due to labor-intensive manual annotation

Engineering Contradiction:

Improveproduct inventory and placement information accuracyVSAvoidtime required for manual annotation

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs self-annotation by automatically detecting products, extracting text from product packaging, and generating audit reports without human intervention. The unsupervised machine learning model processes unannotated images autonomously, eliminating the need for manual annotation while maintaining measurement precision through automated text recognition and product identification algorithms

Inventive Principle:

Principle #25Self-service

2Measurement precision

If annotated training data is used for automatic product inventory auditing, then measurement precision is improved, but loss of substance increases due to labor-intensive manual annotation

Engineering Contradiction:

Improveproduct inventory and placement information accuracyVSAvoidlabor resources for manual annotation

Core Design Contradiction:

Measurement precisionVSLoss of substance

Solution Approach 1:

The system performs self-annotation by automatically detecting products, extracting text from product packaging, and generating audit reports without human intervention. The unsupervised machine learning model processes unannotated images autonomously, eliminating the need for manual annotation while maintaining measurement precision through automated text recognition and product identification algorithms

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical process of manual annotation with an automated computational system. Machine learning models and computer vision algorithms substitute human inspectors, automatically analyzing images, recognizing text, and generating audit reports, thereby eliminating labor resource consumption while maintaining or improving measurement precision

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Ease of operation

If manual inspection is used for audit reports, then ease of operation is maintained, but productivity decreases due to significant time and labor investment

Engineering Contradiction:

Improvesimplicity of manual inspection processVSAvoidaudit report generation speed

Core Design Contradiction:

Ease of operationVSProductivity

Solution Approach 1:

The patent replaces the mechanical process of manual annotation with an automated computational system. Machine learning models and computer vision algorithms substitute human inspectors, automatically analyzing images, recognizing text, and generating audit reports, thereby eliminating labor resource consumption while maintaining or improving measurement precision

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables continuous automated processing of audit reports without interruption by manual tasks. The machine learning model can process multiple images sequentially or in parallel, generating audit reports continuously without the breaks, fatigue, or sequential constraints that limit human inspectors, thereby dramatically improving productivity while maintaining operational simplicity through automated workflows

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentUS11037099B2System and method for live product report generation without annotated training data

Publication Date: 2021.06.15 SALESFORCE INC

AI summary

Embodiments described herein provide a method for obtaining information on product inventory and placement in a retail setting. An image including unannotated image data indicative of the retail setting is received. One or more shelves in the retail setting are determined from the unannotated image data, and the image is segmented into one or more sub-images corresponding to the one or more detected shelves. For each sub-image corresponding to a respective detected shelf, a product name is then and a number of appearances of the product name are detected using text recognition on the respective sub-image. Product inventory information and first placement information are derived based at least in part on the detected number of appearances and a shelf level corresponding to the sub-image.