Material Identification via Basis Function Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radiation-detection methods struggle to differentiate materials with similar densities or compositions, leading to difficulties in identifying potential security threats, as they rely on density and path length, which can result in false positives or negatives.
Innovation Solution
A non-invasive imaging method using a control circuit with multiple models and feasibility criteria to identify materials by producing multiple images with different modalities or spectral channels, employing a small set of basis functions to represent materials as weighted sums, and utilizing singular value decomposition to reduce computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple basis functions (e.g., 95 functions for 95 elements) are used to accurately identify all potential materials, then material identification accuracy is improved, but processing time and computational resources increase unduly
Solution Approach 1:
The patent extracts only the most significant basis functions from the full set of 95 elements. By performing singular value decomposition and selecting top-k basis functions (where k << 95), the system extracts the essential material discrimination capability while removing redundant computational burden. This extraction principle resolves the contradiction by maintaining sufficient identification accuracy with a reduced subset of basis functions.
Solution Approach 2:
The patent applies partial action by using a limited number of top basis functions rather than the complete set. The singular value decomposition identifies which basis functions contribute most to material differentiation, allowing the system to use only the necessary portion (top-k) to achieve effective material identification without the excessive computational cost of processing all 95 basis functions.
2Productivity
If a small number of basis functions are used to reduce processing time and computational resources, then processing efficiency is improved, but material identification accuracy deteriorates
Solution Approach 1:
The patent transforms the basis function representation by applying singular value decomposition to identify the optimal subset of basis functions. This parameter transformation reorganizes the 95-element basis set into a new coordinate system where the first k components capture the most variance in material properties. By changing the representation parameters in this way, the system achieves high processing efficiency with reduced basis functions while preserving identification accuracy.
3Illumination intensity
If radiation-detection views based on density and path length are used, then imaging capability is improved, but material discrimination capability deteriorates when densities are similar
Solution Approach 1:
The patent moves from a single-dimensional density-based representation to a multi-dimensional basis function space. By expressing materials as combinations of 95 elemental basis functions (or a reduced subset), the system adds dimensional complexity that enables discrimination between materials with similar densities. Each material gains a unique fingerprint in this expanded basis space, resolving the limitation of traditional radiography.
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 allows for accurate material identification with significantly reduced computational power, enabling reliable discrimination of various materials using a small number of basis functions, even in challenging settings, by leveraging non-invasive imaging equipment.
Implementation Method 1
The capture of radiation-detection views of a given object using penetrating energy (such as X-rays or the like) is well known in the art. Such radiation-detection views often comprise images having areas that are relatively darker or lighter (or which otherwise contrast with respect to one another) as a function of the density, path length, and/or composition of the constituent materials
Data Source
AI summary
A control circuit having access to information regarding a plurality of models for different materials along with feasibility criteria processes imaging information for an object (as provided, for example, by a non-invasive imaging apparatus) to facilitate identifying the materials as comprise that object by using the plurality of models to identify candidate materials for portions of the imaging information and then using the feasibility criteria to reduce the candidate materials by avoiding at least one of unlikely materials and combinations of materials to thereby yield useful material-identification information.