Radiographic Image Processing Apparatus Striped Artifact Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Radiographic image capturing with FPD type apparatuses results in streaky unevenness due to noise superimposed on electric charges from scanning lines, which is challenging to remove effectively using existing methods.
Innovation Solution
A radiographic image processing apparatus that divides image data into sections, calculates average values along scanning lines, compresses edge profiles, applies adaptive filtering, and generates correction data to accurately extract and remove streaky unevenness components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional CR apparatus is used, then radiographic image can be captured, but streaky unevenness occurs due to noise from scanning lines
Solution Approach 1:
The patent divides the radiographic image into multiple strip-shaped regions along the scanning line direction, allowing independent processing of each region. This segmentation enables the extraction and removal of streaky unevenness specific to each region while preserving the overall image structure and diagnostic information.
Solution Approach 2:
The patent extracts unevenness components from the captured radiographic image by analyzing the relationship between adjacent regions. The extracted unevenness data is then used to generate correction data that can be subtracted from the original image, effectively removing the harmful streaky artifacts while maintaining image quality.
2Reliability
If unevenness removal processing is applied, then streaky unevenness can be removed, but processing time increases
Solution Approach 1:
By dividing the image into strip-shaped regions, the patent can process each region independently using efficient algorithms. This segmentation allows parallel processing and reduces the computational complexity overall, making the unevenness removal faster compared to processing the entire image as a single unit.
Solution Approach 2:
The patent applies unevenness removal processing selectively to regions where streaky artifacts are detected, rather than uniformly processing the entire image. This partial action approach reduces processing time by focusing computational resources only on areas that require correction.
3Productivity
If simple filtering is used, then processing is fast, but streaky unevenness cannot be accurately removed
Solution Approach 1:
The patent segments the image into multiple regions and analyzes the statistical characteristics of each region to determine the appropriate filtering strength. This allows the system to apply stronger filtering where needed and weaker filtering where the image is clean, achieving both speed and accuracy.
Solution Approach 2:
The patent dynamically adjusts the filtering parameters based on the detected unevenness characteristics in each region. The filtering strength and type are adapted in real-time during processing, allowing the system to optimize between processing speed and removal accuracy for each specific region and image condition.
Data Source
AI summary
A radiographic-image processing apparatus that can accurately remove striated artifacts superimposed onto image data taken by an FPD radiographic imaging device includes: a partitioning section that partitions a region, in which image data taken by a radiographic imaging device is arranged two-dimensionally, into segments; a mean-value computation section that computes mean values from the image data along the same scanning line in each segment; an edge compression section that compresses mean-value differentials in boundary regions between the imaged subject and the surroundings thereof; a filtering section that applies an adaptive filter to the differential-compressed profile; and a correction-data creation section that creates correction data on the basis of each datum in the adaptive-filtered profile. Image data in which noise is removed is generated by subtracting corresponding correction data from the image data.