Image Processing Dynamic Range Extension via Band-Signal Noise Reduction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image processing technologies fail to effectively extend the dynamic range and reduce noise in image signals subjected to gradation conversion processing.

Innovation Solution

An image processing apparatus and method that calculates first and second gradation conversion characteristics for different areas of image signals, performs gradation conversion, generates band signals through multi-resolution processing, and applies noise reduction using the second gradation conversion characteristics to combine noise-reduced band signals.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Illumination intensity

If gradation conversion processing is performed to extend dynamic range, then the dynamic range is extended, but noise is amplified in the converted image signals

Engineering Contradiction:

Improvedynamic rangeVSAvoidnoise

Core Design Contradiction:

Illumination intensityVSObject-affected harmful factors

Solution Approach 1:

The image signal is divided into multiple frequency bands using multi-resolution conversion processing. Each band signal is processed independently for noise reduction while preserving the extended dynamic range characteristics. This segmentation allows selective noise reduction in different frequency regions without compromising the overall dynamic range extension.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different noise reduction characteristics are applied to different frequency bands based on their specific properties. The second gradation conversion characteristics are calculated specifically for each band signal, enabling localized optimization of noise reduction while maintaining the global dynamic range extension achieved by the first gradation conversion.

Inventive Principle:

Principle #3Local quality

2Object-affected harmful factors

If noise reduction processing is applied to the entire image signal, then noise is reduced, but the extended dynamic range characteristics are degraded

Engineering Contradiction:

ImprovenoiseVSAvoiddynamic range

Core Design Contradiction:

Object-affected harmful factorsVSIllumination intensity

Solution Approach 1:

The image signal is segmented into multiple frequency bands, allowing noise reduction to be applied selectively to each band rather than uniformly to the entire signal. This preserves the frequency-dependent characteristics introduced by gradation conversion, particularly in the luminance component where dynamic range extension is most critical.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The noise reduction processing uses second gradation conversion characteristics that are specifically calculated for each frequency band, changing the processing parameters according to the frequency band and its specific noise characteristics. This allows adaptive noise reduction that maintains the extended dynamic range in critical frequency regions.

Inventive Principle:

Principle #35Parameter changes

3Device complexity

If uniform noise reduction is applied to all areas of the image, then processing is simplified, but noise reduction effectiveness is reduced in different image areas

Engineering Contradiction:

Improveprocessing complexityVSAvoidnoise reduction effectiveness

Core Design Contradiction:

Device complexityVSObject-affected harmful factors

Solution Approach 1:

The image area is divided into multiple regions, and the first gradation conversion characteristics are calculated separately for each region. This regional segmentation enables the system to adapt to local image characteristics while maintaining manageable processing complexity through systematic calculation methods.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different gradation conversion characteristics are applied to different image regions based on their specific properties. The noise reduction processing also uses region-specific second gradation conversion characteristics, ensuring that each area receives optimized processing tailored to its local characteristics rather than a uniform approach.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS8659684B2Image processing apparatus, program, method and image capturing system

Publication Date: 2014.02.25 OLYMPUS CORPORATION(JP)

AI summary

An image processing apparatus with extended dynamic range includes a first gradation-conversion-characteristics calculating unit that calculates first gradation conversion characteristics for each of a plurality of areas of input image signals; and a gradation conversion unit that performs gradation conversion on each of the areas of the image signals by using the first gradation conversion characteristics. A filter processing and reduction processing unit generates a plurality of band signals having mutually different frequency bands by multi-resolution conversion processing from the image signals subjected to the gradation conversion. A second gradation-conversion-characteristics calculating unit calculates second gradation conversion characteristics associated with the individual band signals from the first gradation conversion characteristics An NR processing unit performs processing for noise reduction on the individual band signals by using the second gradation conversion characteristics; and an adding unit combines the band signals subjected to the processing for noise reduction.