Ultrasonic Image Processing Using Motion Vector Scalar Conversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional ultrasonic image processing methods fail to accurately reflect vector information and degrade border sharpness when using low-pass filters for noise removal, and are not adaptable to varying analysis techniques based on regions or diagnostic purposes.
Innovation Solution
Converting motion vectors into scalar distributions using eigenvalue expansion and applying a similarity filter to each scalar distribution for noise removal while preserving edge information, allowing for adaptable analysis techniques based on regions of interest.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a uniformly smoothing type low-pass filter is used to remove noise from motion vector distribution, then noise is removed, but the sharpness of tissue borders is degraded
Solution Approach 1:
The patent applies different filtering strategies to different regions of the motion vector distribution. A similarity filter is used to preserve edge information in regions with significant motion changes, while a low-pass filter is applied to smooth regions with uniform motion patterns. This local differentiation allows noise removal without degrading border sharpness.
Solution Approach 2:
The patent segments the motion vector distribution into multiple regions based on motion characteristics. By dividing the image into regions with different motion patterns (e.g., high motion variance vs. low motion variance areas), the system can apply appropriate filtering operations to each segment independently, preserving edges while removing noise.
2Device complexity
If conventional scalar distribution methods are used to convert motion vectors, then processing is simplified, but vector information is not accurately reflected
Solution Approach 1:
The patent transforms the motion vector distribution from a two-dimensional vector field into a scalar distribution by calculating motion energy or divergence at each pixel. This dimensional transformation simplifies the data structure while preserving essential motion information through mathematical operations that maintain the physical meaning of the vectors.
Solution Approach 2:
The patent changes the representation parameters of motion vectors from complex vector components (magnitude and direction) to scalar parameters such as motion energy, divergence, or rotation. This parameter transformation simplifies further processing while accurately reflecting the essential characteristics of the motion field through appropriate mathematical formulations.
Data Source
AI summary
An ultrasonic image processing method and device, and an ultrasonic image processing program which can correspond to analytical methods different depending on a region or the purpose of a diagnosis or treatment. The ultrasonic image processing method comprises an image data creation step which stores a detection result obtained by irradiating a subject with ultrasonic waves by an irradiating section and detecting an ultrasonic signal from the subject by a detecting section and creates at least two-frame image data different in detection timing on the basis of the stored detection result, a motion vector distribution image creation step which creates a motion vector distribution image on the basis of a predetermined motion vector analysis through the use of a plurality frames of the image data, and a conversion step which converts a vector distribution image to a scalar distribution image on the basis of a plurality of established regions of interest (ROI).