Image Region Classification Using Geographic Location Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for classifying regions in digital images struggle to accurately determine specific materials and objects, especially when contextual information is lacking, leading to incorrect results due to similar color and texture characteristics among materials like snow and clouds, and fail to distinguish between indoor and outdoor images based on location information alone.

Innovation Solution

A digital image capture device that utilizes geographic location information to generate a material belief map by integrating location data with image processing, employing a probabilistic graphical model to refine region classification through spatial context and prior probabilities specific to the capture location, thereby enhancing image classification accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If image region classification is performed using only image features (color and texture), then the classification process is simple, but the accuracy is poor due to similar characteristics among different materials

Engineering Contradiction:
Improveclassification accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent combines multiple information sources (image features, geographic location data, and contextual scene information) into a unified classification system. The region classification module integrates visual data with external context data to produce more accurate material classification, resolving the contradiction by merging simple image processing with additional data layers without creating a complex distributed system

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces an intermediary contextual layer (geographic location and scene type) that mediates between raw image features and final classification. This intermediary provides disambiguation for materials with similar visual characteristics by incorporating location-based prior probabilities, improving accuracy without requiring direct complex interactions between multiple processing modules

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If geographic location information is used to improve region classification, then classification accuracy improves, but the system complexity increases due to integration of additional data sources

Engineering Contradiction:
Improveregion classification accuracyVSAvoiddata integration complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary actions by pre-computing and storing contextual information (geographic location data, scene type probabilities, and material priors) before the actual classification task. This allows the classification module to quickly access pre-processed contextual data without performing complex real-time computations, improving accuracy while managing system complexity through advance preparation

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements a universal contextual processing framework that handles multiple data types (geographic coordinates, scene classification, time information) through a single integrated architecture. The belief map generation and updating mechanism serves multiple functions: it processes different material classes, integrates various contextual sources, and adapts to different locations, reducing overall system complexity through multi-functionality

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Measurement precision

If location-based contextual information is integrated into image classification, then the ability to distinguish materials with similar characteristics improves, but the processing time increases

Engineering Contradiction:
Improvematerial distinction accuracyVSAvoidclassification processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies partial action by selectively updating belief maps only for regions where contextual information provides disambiguation value. Rather than processing every pixel with full contextual analysis, the system focuses computational resources on ambiguous regions identified through initial feature analysis, improving material distinction while reducing overall processing time through selective application

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP2033140B1Classifying image regions based on picture location
Publication Date: 2012.01.11 EASTMAN KODAK CO
  • EP2033140B1 patent drawingFigure 1
  • EP2033140B1 patent drawingFigure 2
  • EP2033140B1 patent drawingFigure 3

AI summary

A method of classifying regions in a digital image or video captured by an image capture device includes providing a geographic location determining device associated with the image capture device; using the location determining device to measure the image capture location at substantially the time that the digital image or video was captured; and classifying regions in the captured digital image or video into one or more classes based on the image-capture location.