Method and System for Determining Local Quality of Surface Data Extracted from Volume Data

Simple SummaryContent extracted from patent full text and abstract with AI.

This patent describes a method and system for evaluating the local quality of surfaces extracted from three-dimensional (3D) volume data, such as those generated by computed tomography (CT), magnetic resonance imaging (MRI), or ultrasound tomography. For each surface point identified by existing surface extraction methods, the invention analyzes the surrounding volume data and computes a quality metric based on features like the sharpness, contrast, noise, and similarity to reference profiles of the grayscale data in that region. The resulting quality score is attached to the surface point, helping users understand the reliability or precision of specific regions of the extracted surface model. The system can visualize these quality scores using color coding, guide further analysis, or improve fitting processes and data fusion.

Use CasesContent extracted from patent full text and abstract with AI.

• Quality assessment of 3D surface models generated from industrial CT scans for dimensional metrology or defect analysis.
• Enhancing reliability in quality control inspections in manufacturing, especially for complex internal structures.
• Optimizing the settings or orientation of tomographic scans based on objective surface quality metrics, reducing operator dependency.
• Automated comparison and alignment of measured objects with CAD models or known geometries, weighting regions by reliability.
• Fusion or combination of multiple scan datasets, prioritizing high-quality regions for superior composite models.
• Supporting medical imaging analysis (MRI, CT, ultrasound) by identifying areas with reliable anatomical boundaries or pathology.
• Benchmarking and optimizing surface extraction algorithms based on local quality metrics, even without reference data.

BenefitsContent extracted from patent full text and abstract with AI.

• Provides an objective, local quality metric for each point on a 3D surface, improving trustworthiness in surface-based measurements.
• Highlights unreliable or artifact-prone surface regions, allowing users to focus on accurate data and avoid errors in downstream tasks.
• Enables automated visualization of surface quality, making it easy to communicate and understand measurement confidence.
• Facilitates better alignment and fitting in metrology, by weighting or excluding low-quality areas when comparing with models or fitting geometric primitives.
• Improves process optimization: scan parameters and object positioning can be systematically refined for best measurement quality in regions of interest.
• Supports data fusion, allowing the creation of composite models that maximize high-quality surface information from multiple scans.
• Reduces dependence on user interpretation and prior knowledge, which is crucial for complex or inaccessible objects and for automation.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

The aim of the invention is to determine the local quality of surface data (O) extracted from a volume data set (V) by means of a surface determination method. An environment in the volume data set (V) is determined for each surface point of the surface data (O). Using the curve of the grayscale values of voxels from said environment, at least one quality characteristic (Q) is derived which characterizes the quality of the respective examined surface point. The quality characteristic (Q) or each quality characteristic is output together with coordinates of the respective examined surface point as the method result (O').