Visual Object Detection in Video Content for E-commerce

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current technologies lack the ability to allow viewers to express interest in or obtain information about objects or products viewed in videos or images without keyword searches or scanning QR codes, limiting engagement and sales opportunities in e-commerce and Internet marketing.

Innovation Solution

A method that segments video content, identifies objects using convolutional neural networks, and matches them with products in a database, enabling viewers to receive information or advertisements about the objects through a networked system accessible via various devices.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If visual search is used to allow users to interact with objects in video content, then user engagement and information access are improved, but the ability to capture and process visual data for metadata generation is lost

Engineering Contradiction:

Improveuser interaction with video objectsVSAvoidmetadata capture from visual content

Core Design Contradiction:

Ease of operationVSLoss of information

Solution Approach 1:

The system performs preliminary analysis of video content by segmenting it into meaningful sections and pre-identifying objects within those segments before user interaction occurs. This allows the system to have metadata and object information ready when users perform visual searches, eliminating the loss of information while maintaining ease of interaction.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary visual search system that bridges between video content and user queries. This intermediary captures visual data from video segments, processes it through object recognition algorithms, and generates metadata that enables both user interaction and information retention simultaneously.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If video content is segmented and analyzed for object detection, then product matching and targeted advertising are improved, but processing time and computational resources increase

Engineering Contradiction:

Improveproduct matching efficiencyVSAvoidvideo processing time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The patent divides video content into distinct segments based on visual content analysis, allowing object detection to be performed on smaller, manageable portions rather than entire videos. This segmentation enables parallel processing and reduces overall computation time while maintaining comprehensive product matching capabilities across the full video content.

Inventive Principle:

Principle #1Segmentation

3Device complexity

If conventional search methods are used for product information, then system simplicity is maintained, but user engagement and sales opportunities are limited

Engineering Contradiction:

Improvesearch system simplicityVSAvoidsales opportunity generation

Core Design Contradiction:

Device complexityVSProductivity

Solution Approach 1:

The patent replaces conventional text-based keyword search mechanisms with visual search capabilities using image recognition technology. This substitution allows users to interact with video content more naturally by capturing and analyzing visual elements, thereby increasing engagement and sales opportunities without significantly complicating the user interface.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS10769444B2Object detection from visual search queries

Publication Date: 2020.09.08 GOH SOO SIAH

AI summary

This invention includes a system and method of populating a data-base with known objects. The database can be populated with off-line data augmentation (e. g. a web crawler) or by aligning known objects and metadata clusters with defined content. A viewer can query images from live or offline media. Objects in the viewers query are linked with similar objects or recommended products in the database.