Video-Based Product Shelf Analysis for Unpurchased Interest Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing product analysis systems struggle to accurately determine customer interest in products that were not purchased, as they require costly equipment and tags, and cannot differentiate between product retrieval and placement actions.

Innovation Solution

A product analysis system that includes a detection mechanism for identifying changes in product shelves from video footage, a classification mechanism to categorize these changes, and a specification mechanism to determine the frequency of customer interest in unpurchased products based on the classified changes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If sensors or tags are introduced to detect customer behavior, then measurement precision of customer interest is improved, but device complexity and cost increase

Engineering Contradiction:

Improvecustomer interest detection accuracyVSAvoidequipment and tags

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent uses image capture devices to create visual copies of customer interactions with products. Instead of physical sensors on products, the system captures images of customers reaching for or returning products and processes these images to detect interest behaviors, thereby avoiding the need for complex physical tagging systems

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces mechanical sensor systems with an optical-based image processing system. By using cameras to capture and analyze visual data of customer behaviors, the system substitutes complex mechanical detection devices with simpler optical equipment and computational algorithms

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If hand-reaching detection is used to measure customer interest, then productivity of interest analysis is improved, but measurement precision deteriorates due to inability to differentiate retrieval from placement actions

Engineering Contradiction:

Improveinterest analysis speedVSAvoidbehavior differentiation accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent segments the hand-reaching detection process into multiple distinct analysis stages: initial image capture, change detection to identify reaching actions, classification to differentiate between retrieval and placement, and final interest determination. This segmentation allows each stage to focus on specific aspects, improving overall precision while maintaining efficiency

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system incorporates feedback mechanisms where the image processing results are continuously refined. By analyzing sequences of images and comparing changes over time, the system adjusts its detection algorithms to better distinguish between retrieval and placement actions, improving measurement precision through iterative refinement

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS11443503B2Product analysis system, product analysis method, and product analysis program

Publication Date: 2022.09.13 NEC CORP

AI summary

A product analysis system 800 includes a detection means 810, a classification means 820, and a specification means 830. The detection means 810 detects an area of change in a product shelf from a video of the product shelf. The classification means 820 classifies the change in the product shelf in the detected area of change. The specification means 830 specifies the frequency at which a customer was interested in but did not purchase a product on the basis of the classification of the change.