Localized CT Value Distribution Analysis for Threat Item Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated object recognition systems in computed tomography (CT) face challenges in accurately identifying threat items concealed within electronic devices due to image artifacts, which distort density and atomic number values, leading to false negatives.
Innovation Solution
A method that involves segmenting a three-dimensional image of an object into sub-regions, binning voxels based on CT values, and classifying items as potential threats when a specified threshold of voxels within a sub-region match known CT values for a threat type, thereby overcoming the issue of artifacts and non-contiguous threat material.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated object recognition systems use traditional image analysis methods to identify threat items, then the system can process images quickly, but image artifacts from high density components distort CT values and lead to false negatives
Solution Approach 1:
The patent divides the three-dimensional image into multiple sub-regions and further segments each sub-region into voxels. By analyzing CT value distributions within each voxel and comparing them to reference CT values for different material types, the system can identify threat items even when artifacts are present. This segmentation approach allows the system to process images efficiently while maintaining high identification accuracy by localizing analysis to small, manageable regions.
Solution Approach 2:
The patent applies local quality analysis by examining CT value distributions within individual voxels and sub-regions rather than analyzing the entire image globally. Each voxel's CT values are compared against reference ranges for specific material types (e.g., explosives, electronics). This localized approach enables the system to identify threat items with high accuracy while maintaining processing speed, as artifacts in one region do not compromise the analysis of other regions.
2Device complexity
If the system analyzes the entire three-dimensional image as a single region, then the analysis is simpler, but it cannot detect non-contiguous threat material hidden within electronic devices
Solution Approach 1:
The patent segments the three-dimensional image into multiple sub-regions and further divides each sub-region into individual voxels. This hierarchical segmentation allows the system to detect non-contiguous threat material by identifying clusters of voxels with matching CT values, even when those clusters are separated by electronic device components. The segmentation increases detection precision while keeping analysis complexity manageable through automated processing.
Solution Approach 2:
The patent transitions from two-dimensional image analysis to three-dimensional volumetric analysis by dividing the image into voxels (volume elements). This dimensional enhancement allows the system to detect threat material that may be concealed within the three-dimensional structure of electronic devices, improving detection precision by analyzing spatial relationships and CT value distributions throughout the entire volume rather than just surface features.
3Measurement precision
If the system uses artifact reduction techniques to improve image quality, then measurement accuracy improves, but processing time and computational resources increase significantly
Solution Approach 1:
The patent extracts and analyzes only the relevant CT value data from each voxel, comparing measured CT values directly against reference CT value ranges for different material types. By extracting only the essential measurement data (CT values) and comparing them to pre-established reference ranges, the system achieves high measurement accuracy without requiring time-consuming artifact reduction processing. The method takes out the critical information needed for identification while discarding unnecessary processing steps.
Solution Approach 2:
The patent changes the approach from attempting to correct artifact-distorted CT values to instead comparing the distorted CT values against reference ranges that account for expected variations. By changing the parameter comparison strategy from seeking perfect accuracy to accepting range-based matching, the system maintains measurement precision while avoiding time-consuming artifact reduction techniques. The system evaluates whether CT values fall within acceptable ranges for specific material types rather than attempting to eliminate all sources of variation.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach improves the accuracy of identifying concealed threat items by allowing classification based on localized CT value distributions, even when threat material is visually or physically non-contiguous, reducing false negatives and enhancing security screening efficiency.
Implementation Method 1
a radiation imaging modality is utilized to facilitate baggage screening. For example, a CT system may be used to provide security personnel with two and/or three dimensional views of objects
Implementation Method 2
a CT system may be used to provide security personnel with two and/or three dimensional views of objects
Data Source
AI summary
Among other things, one or more systems and/or techniques for classifying an item disposed within an object are provided herein. A three-dimensional image of the object (e.g., a bag) is segmented into a set of item representations (e.g., laptop, thermos, etc.). An item is identified from the set of item representations based upon item features of the item, such as the laptop that could be used to conceal an item of interest such as an explosive. A region comprising a three-dimensional image of the item is divided into a set of sub-regions (e.g., a first sub-region encompassing a screen, a second sub-region encompassing a motherboard, etc.). The item is classified as a potential first type of item (e.g., an explosive laptop) when any sub-region has a number of voxels, with computed tomography (CT) values within a range of known CT values for a first type of item, exceeding a threshold.