3D Image Processing Apparatus Noise Reduction via Search Area Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing techniques face inefficiencies in calculating the voxel with closest luminance from the neighborhood of a corresponding voxel, leading to noise in subtraction images due to errors in registration or intensity differences.

Innovation Solution

An image processing apparatus sets a search area around the corresponding position in the second image, using a shift of voxel position to determine the search area size based on voxel sizes of both images, and compares intensities within this area to obtain a difference value, reducing noise in three-dimensional subtraction images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a voxel-matching method is used to reduce noise in subtraction images, then noise reduction is achieved, but calculation efficiency deteriorates due to the complexity of finding the voxel with closest luminance from the neighborhood

Engineering Contradiction:
Improvenoise reduction in subtraction imageVSAvoidcalculation efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent divides the search area into multiple regions and processes voxels in segments rather than exhaustively searching the entire neighborhood. This segmentation approach reduces the computational complexity while maintaining effective noise reduction by focusing on the most relevant voxel candidates.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Instead of examining all voxels in the neighborhood (excessive action), the patent applies a threshold-based selection to consider only partial voxels that meet certain luminance criteria. This partial action approach maintains sufficient noise reduction effectiveness while significantly improving calculation efficiency.

Inventive Principle:
Principle #16Partial or excessive action

2Reliability

If the search area is enlarged to find the closest luminance voxel, then noise reduction improves, but calculation time increases

Engineering Contradiction:
Improvenoise reductionVSAvoidcalculation time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary sorting or pre-processing of voxel luminance values before the actual search. By preparing the data structure in advance (e.g., sorting voxels by luminance or creating a hierarchical representation), the system can quickly identify the closest luminance match without examining all voxels, thus reducing calculation time while maintaining an adequate search area for noise reduction.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3662835B1Image processing device, image processing method, image processing system, and program
Publication Date: 2024.12.04 CANON KK
  • EP3662835B1 patent drawingFigure 1
  • EP3662835B1 patent drawingFigure 2
  • EP3662835B1 patent drawingFigure 3

AI summary

An apparatus that can obtain a subtraction image efficiently is provided. An image processing apparatus obtains a target image constituted by a set of voxels arranged in a discretized manner; sets a search area in the target image; and obtains, in a partial area included in the search area, on the basis of at least one of a voxel value included in the partial area or an interpolated value obtained by interpolation of a voxel of the target image, at least one of a maximum and a minimum of a voxel value or an interpolated value within the search area.