Radiation Image Processing Apparatus Frequency Band Decomposition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques for radiation images do not effectively enhance frequency components according to the object composition or user preference, often resulting in unsuitable images with noise or insufficient contrast.
Innovation Solution
An image processing apparatus that decomposes input image data into band-limited signals, selects preset data with frequency-response tables to perform nonlinear conversions, and reconstructs enhanced images by adding the converted signals to the original data, allowing for customizable frequency enhancement based on object composition or user preference.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frequency enhancement is applied to radiation images, then image contrast and interpretability are improved, but noise and artifacts are also enhanced
Solution Approach 1:
The image processing is segmented into multiple frequency bands using wavelet transformation. Different enhancement strategies are applied to different frequency components: high-frequency components (containing edge information) are enhanced with caution to avoid noise amplification, while low-frequency components (containing overall structure) are enhanced more aggressively. This segmentation allows selective noise suppression while maintaining contrast enhancement benefits.
Solution Approach 2:
The enhancement process applies different processing characteristics to different regions of the image based on local properties. In regions with high signal-to-noise ratio, stronger enhancement is applied. In regions with low signal-to-noise ratio or near edges, enhancement is moderated to prevent artifact generation. This local adaptation resolves the contradiction by making enhancement quality spatially variable.
2Measurement precision
If strong frequency enhancement is applied, then image interpretability improves, but artifacts near edges are generated
Solution Approach 1:
The enhancement coefficient is made dynamic rather than fixed. It varies based on local image characteristics such as gradient magnitude, noise level, and frequency content. This dynamic adjustment allows strong enhancement where it improves interpretability without generating artifacts, and reduces enhancement where artifacts would compromise reliability.
Solution Approach 2:
The processing incorporates feedback mechanisms where the output of one processing stage informs subsequent stages. Artifact detection algorithms monitor the enhanced image and feed back to adjust enhancement parameters in real-time, suppressing enhancement in regions where artifacts are detected while maintaining enhancement in clean regions.
3Ease of manufacture
If frequency enhancement is applied uniformly, then processing simplicity is maintained, but results do not match object composition or user preference
Solution Approach 1:
The processing system is designed with multi-functionality to handle different object types and user preferences. Multiple preset enhancement profiles are provided (e.g., bone-enhanced, soft-tissue enhanced, edge-enhanced) that can be selected based on the imaging application. The underlying algorithm remains universal, but its parameters are adapted to different requirements, resolving the contradiction between simplicity and adaptability.
Data Source
AI summary
An image processing apparatus includes the following. A hardware processor decomposes a signal value of input image data into band-limited signals having different frequency bands from each other. A storage stores pieces of preset data. Each of the pieces of preset data comprises tables to associate frequency with a response and to prescribe different response properties from each other. The hardware processor selects a piece of preset data from the pieces of preset data stored in the storage, converts the decomposed band-limited signals on a basis of tables in the selected piece of preset data, reconstructs the converted band-limited signals into enhanced image data, and generates a frequency-enhanced image through addition of the enhanced image data which is multiplied by a predetermined enhancement coefficient to the input image data.