Modeling realistic virtual objects within a high-throughput x-ray simulation framework

Coccarelli, David; Gehm, Michael E.; Greenberg, Joel A.

doi:10.1117/12.2558947

Cited by 4 publications

(6 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To overcome these limitations, we have previously described our novel, computational imaging approach to XRDI, which allows for full-tunnel imaging of baggage at conventional belt speeds. 5 Our work in this paper focuses on how X-ray diffraction imaging can be coupled with existing X-ray transmission measurements 6, 7 to produce a system with a lower false alarm rate than either system alone due to its superior material characterization capabilities. 8,9 Specifically we provide detail on the utility of high quality XRD reconstructions in discriminating threat materials from benign objects.…”

Section: Introductionmentioning

confidence: 99%

Improved reconstruction and classification performance in a hybrid CT+XRD explosives detection system

Richmond,

Pike,

Franco

et al. 2024

Anomaly Detection and Imaging With X-Rays (ADIX) IX

View full text Add to dashboard Cite

Identifying and intercepting prohibited items and explosives is a critical focus of aviation security. While computed tomography (CT) systems represent the industry standard for detecting explosives in baggage, X-ray diffraction imaging (XRDI) systems have shown increasing performance and commercial viability. Our approach to explosives detection involves the combination of CT and XRDI into a single, hybrid system where both the CT and XRDI data are utilized in the reconstruction and classification algorithms. In this work, we focus on comparing multiple reconstruction and classifier implementations and quantifying the resulting performance. Our analysis shows higher quality reconstructions lead to improved material separability, better classification performance (detection and false alarm rates), and reduces model uncertainty. Through this work, we demonstrate the relationship between improved quality of reconstructions and the separability of threat from non-threat objects in the domain of explosives detection.

show abstract

Section: Introductionmentioning

confidence: 99%

Improved reconstruction and classification performance in a hybrid CT+XRD explosives detection system

Richmond,

Pike,

Franco

et al. 2024

Anomaly Detection and Imaging With X-Rays (ADIX) IX

View full text Add to dashboard Cite

show abstract

“…Recent work in synthetic data generation offers the potential to dramatically reduce the amount and types of empirical data collection required, while simultaneously enabling more rapid and successful algorithm development as well as system testing and evaluation [1,2]. There are, however, different types of synthetic data, including methods that add representative examples of items or scenarios based on training with empirical data (e.g., using a generative adversarial network, or GAN [3]), methods that numerically augment empirical measurements (e.g., data augmentation techniques [4]), and fully-virtual methods (e.g., physics-based synthetic data [5][6][7][8]). It is worth noting that only physics-based synthetic data, in which the virtual X-ray scanner and bags exist only as numerical representations, does not rely on an empirical instantiation of a real scanner or bag to produce data.…”

Section: Introductionmentioning

confidence: 99%

“…Over the last few years, we have developed an end-to-end framework for rapidly generating large volumes of high-fidelity, physics-based synthetic X-ray data [7,8]. This framework has undergone successful independent validation in which approximately 2,000 computed tomography (CT) volumes of baggage were generated and compared against empirical data from a deployed, multi-energy CT system.…”

Section: Introductionmentioning

confidence: 99%

“…While we have previously shown that synthetic datasets have the capacity to directly match the statistics of an empirical dataset [6,7,9,10], approaching synthetic data validation and application from this perspective fundamentally limits the value of the synthetic data. In this paper, we discuss the ways in which synthetic datasets differ from empirical datasets and how these differences can result in more robust, interpretable detection algorithms and quantitative, complete assessments of a system's detection performance.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Unique opportunities to sample parameter space using physics-based synthetic x-ray data

Greenberg¹,

Pike²,

Coccarelli³

2023

Anomaly Detection and Imaging With X-Rays (ADIX) VIII

View full text Add to dashboard Cite

Physics-based synthetic data generation has been shown to complement empirical data in the rapid development of robust algorithms as well as their testing and interpretation. Quadridox’s QSim RT framework can generate rapidly large volumes of realistic X-ray data and associated metadata; while this data can be made to match traditional empirical data collections, the true value of physics-based synthetic data resides in the ways that it goes beyond empirical data to fully explore and map out the parameter space in which the data resides.

show abstract

“…To reduce the statistical noise associated with random sampling, these methods generally need to simulate a huge number of photons and therefore require long computation times. Ray-tracing simulations such as QSim [4] compute average quantities for the photon distribution, thereby sacrificing some accuracy to greatly reduce computation time. An advantage of all simulation approaches is that images can be generated before any data are collected and even for x-ray CT systems, which do not yet exist.…”

Section: Introductionmentioning

confidence: 99%

3D threat image projection through dual-energy decomposition

Emerman¹,

DeMasi

Stroker

et al. 2023

Anomaly Detection and Imaging With X-Rays (ADIX) VIII

View full text Add to dashboard Cite

Synthetic data are commonly used to train machine learning models in domains where real data are sparse. In this work, we describe a method to generate synthetic x-ray imaging data by inserting objects into a dual-energy computed tomography scan while simultaneously inserting the beam-hardening and noise artifacts that corrupt real data. This type of data augmentation is useful for training classifiers, for example, by artificially increasing the prevalence of objects of interest in a dataset. This work extends existing 3D Threat Image Projection methods by using dual-energy decomposition to model the energy-dependence of attenuation values in the sinogram data. By summing linear attenuation coefficient functions, objects can be inserted directly into a sinogram while accounting for beam-hardening in the insertion region. In addition, we introduce a calibration method to model the change in noise levels resulting from the insertion of attenuating objects. The performance of the method is demonstrated on a simple phantom scanned with a benchtop microCT system.

show abstract

Modeling realistic virtual objects within a high-throughput x-ray simulation framework

Cited by 4 publications

References 0 publications

Improved reconstruction and classification performance in a hybrid CT+XRD explosives detection system

Improved reconstruction and classification performance in a hybrid CT+XRD explosives detection system

Unique opportunities to sample parameter space using physics-based synthetic x-ray data

3D threat image projection through dual-energy decomposition

Contact Info

Product

Resources

About