Landmark Database Construction via Web Data Mining

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing landmark recognition systems rely on biased databases generated from specific user groups' photo-sharing websites, lacking a comprehensive and broad range of publicly available photos, which limits their ability to recognize a wide variety of landmarks.

Innovation Solution

A computer-implemented method and system that builds a landmark database by mining web data from various sources, including travel websites and image search engines, to generate visual models of landmarks, enabling recognition across a broader range of images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If landmark databases are built from photo-sharing websites, then the database can be automatically generated, but the database becomes biased to particular user group interests and lacks comprehensiveness

Engineering Contradiction:

Improveautomatic database generationVSAvoidcomprehensiveness of landmark coverage

Core Design Contradiction:

Extent of automationVSAdaptability or versatility

Solution Approach 1:

The patent applies universality by using multiple diverse data sources (photo-sharing websites, travel websites, image search engines) instead of relying on a single source. This multi-functional approach ensures the landmark database covers a broader range of landmarks across different user interests and geographic locations, making the system more versatile and comprehensive while maintaining automatic generation capabilities

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Ease of manufacture

If landmark databases are built from a single source, then the database generation process is simple, but the database lacks breadth and covers only a limited range of landmarks

Engineering Contradiction:

Improvesimplicity of database generationVSAvoidnumber of diverse landmarks

Core Design Contradiction:

Ease of manufactureVSQuantity of substance

Solution Approach 1:

The patent merges data from multiple sources including photo-sharing websites, travel websites, and image search engines into a unified landmark database. This combining approach increases the quantity and diversity of landmarks in the database while maintaining a systematic generation process that handles multiple sources efficiently

Inventive Principle:

Principle #5Merging (Combining)

3Loss of time

If visual models are generated from limited photo sources, then the processing time is reduced, but the landmark recognition accuracy decreases for a wide range of landmarks

Engineering Contradiction:

Improvevisual model generation timeVSAvoidlandmark recognition accuracy

Core Design Contradiction:

Loss of timeVSMeasurement precision

Solution Approach 1:

The patent applies preliminary action by pre-generating visual models for a comprehensive set of landmarks from multiple sources before actual recognition tasks. This advance preparation ensures that when landmark recognition is needed, the system can quickly compare images against pre-existing visual models, maintaining both speed and high accuracy across diverse landmarks

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS9323792B2Web mining to build a landmark database and applications thereof

Publication Date: 2016.04.26 GOOGLE LLC

AI summary

This invention relates to building a landmark database from web data. In one embodiment, a computer-implemented method builds a landmark database. Web data including a web page is received from one or more websites via one or more networks. The web data is interpreted using at least one processor to determine landmark data describing a landmark. At least a portion of the landmark data identifies a landmark. Finally, a visual model is generated using the landmark data. A computing device is able to recognize the landmark in an image based on the visual model.