Hierarchical Image Blur Detection Using Edge and Frequency Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional image blur estimation techniques are inefficient and produce unreliable results due to their universal nature, failing to account for specific features of subjects like faces, leading to difficulties in accurately determining the degree and direction of blur.

Innovation Solution

A hierarchical approach combining edge analysis and frequency analysis to classify images, where edge analysis quickly identifies blurry images and frequency analysis further processes them to determine the degree and direction of blur, utilizing preprocessing to normalize images and reduce noise.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional universal blur estimation techniques are used, then processing is simpler, but measurement precision deteriorates due to failure to account for specific subject features

Engineering Contradiction:
Improveblur detection accuracyVSAvoidprocessing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The blur detection process is segmented into two distinct stages: edge analysis for initial classification and frequency analysis for detailed evaluation. This segmentation allows the system to apply different processing depths to different images, achieving high precision for problematic cases while maintaining efficiency for clear images.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system changes processing parameters dynamically based on image characteristics. Edge analysis serves as a preliminary parameter check that determines whether full frequency analysis is necessary, allowing the system to adapt its measurement precision to the actual blur level present in each image.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If edge analysis alone is used, then processing time is reduced, but measurement precision deteriorates due to scale-dependent edge width variations

Engineering Contradiction:
Improveprocessing speedVSAvoidblur detection accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

Edge analysis is performed as a preliminary action before frequency analysis. This preliminary step quickly identifies images that are clearly sharp or clearly blurry, allowing the system to avoid unnecessary complex processing for obvious cases while ensuring thorough analysis for ambiguous cases.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

Edge analysis serves as an intermediary step between simple rejection and full frequency analysis. It acts as a mediator that filters images based on initial assessment, directing only those requiring detailed evaluation to the more computationally intensive frequency analysis stage.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If frequency analysis is applied to all images, then measurement precision is improved, but loss of time increases due to computational intensity

Engineering Contradiction:
Improveblur detection accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system applies partial action by performing frequency analysis only on images that require it, rather than applying full frequency analysis to all images. This selective approach ensures measurement precision is improved for problematic images while avoiding unnecessary time loss for images that can be adequately assessed through edge analysis alone.

Inventive Principle:
Principle #16Partial or excessive action

4Reliability

If universal blur estimation is used, then device complexity is reduced, but reliability deteriorates due to inability to handle different subject classes

Engineering Contradiction:
Improveblur detection reliabilityVSAvoidprocessing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The hierarchical processing system provides universal applicability across different image types and blur conditions. By combining edge analysis and frequency analysis in a unified framework, the system reliably handles diverse subject classes and blur scenarios without requiring separate specialized processing paths for each case.

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

Data Source

PatentUS9361672B2Image blur detection
Publication Date: 2016.06.07 VIEWDLE INC

AI summary

A system of estimating image blur in digital images stored in a digital file. The disclosed system may calculate a blur value and direction of motion causing image blur based on detected edges within an input image and spectral energy information of the input image in the frequency domain.