Breast Radiography Processing for Noise-Reduced Calcification Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques for detecting calcification lesions in breast radiography images are hindered by noise, particularly from differences in mammary gland distribution and breast thickness, making it difficult to identify small calcifications.
Innovation Solution
An image processing apparatus that normalizes the rate of change in brightness gradient using a variation degree, such as standard deviation, and applies threshold value processing to detect calcification lesions by reducing the influence of noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image analysis is performed using density gradient of radiation image, then calcification lesions can be detected, but noise from mammary gland distribution and breast thickness differences significantly affects detection accuracy
Solution Approach 1:
The patent changes the parameter used for detection from simple density gradient to normalized rate of change of brightness gradient. This parameter transformation allows the system to account for variations in mammary gland distribution and breast thickness by normalizing against local variation degrees, thereby improving detection accuracy while reducing noise influence.
Solution Approach 2:
The patent performs preliminary normalization processing on the brightness gradient before detection. By calculating the variation degree in each local region and normalizing the rate of change accordingly, the system prepares the image data in advance to reduce noise effects, making the subsequent detection more accurate.
2Measurement precision
If synthesized two-dimensional image is generated by synthesizing multiple tomographic images, then small calcification lesions become more visible, but noise from different regions is also amplified
Solution Approach 1:
The patent applies normalization to the rate of change of brightness gradient before synthesizing tomographic images. This parameter transformation ensures that when multiple images are synthesized, the noise from different regions is not simply amplified but is instead normalized relative to local variation, maintaining visibility of small calcifications while controlling noise.
Solution Approach 2:
The normalized rate of change of brightness gradient acts as an intermediary parameter between the raw tomographic images and the final synthesized image. This intermediary processing step transforms the data in a way that preserves lesion visibility while reducing noise impact before the final synthesis occurs.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
Provided are an image processing apparatus, a radiography system, and a program that can detect a calcification lesion from a radiation image of a breast in a state in which an influence of noise is reduced, as compared with the related art. An image processing apparatus includes an image acquisition unit (30) that acquires a radiation image obtained by imaging a breast with radiation, a normalization unit (32) that normalizes a rate of change in a brightness gradient in the acquired radiation image using a variation degree that indicates a degree of variation in a local region for each local region, and a calcification detection unit (33) that detects a position of a calcification lesion in the radiation image by performing threshold value processing using a predetermined threshold value on the normalized rate of change in the brightness gradient.