This PDF file contains the front matter associated with SPIE Proceedings Volume 12104, including the Title Page, Copyright information, Table of Contents, and Conference Committee listings.

The Transportation Security Laboratory (TSL) routinely performs test and evaluation of explosives detection systems using real explosives, and detection algorithms are commonly trained using data acquired using real explosives. However, in some cases, explosives are either too dangerous to handle or otherwise restricted, in which case inert simulants are used. Simulants are developed to mimic the physical properties of explosives and other hazardous materials to eliminate the safety risks associated with testing the real threat. The simulant development process at the TSL involves combining chemicals so that the physical properties of the resulting formulation match the target threat and the formulation remains stable and inert. This process becomes increasingly difficult as the number of targeted physical properties increases. To facilitate simulant development, a MATLAB-based program was developed to generate simulant formulas by optimizing the mass percentage of the ingredients such that, when combined, the mixture exhibits the same physical properties as the selected target. In this study, five powder simulants were optimized, manufactured, and tested, and the measured properties were compared to the theoretical values generated by the simulant development program. The results demonstrated the program’s accuracy at predicting each formulation’s physical properties. The accuracies ranged from 79 to 100 percent, with lower accuracies being influenced by difficulties in predicting the formulation’s packing density. The simulant development tool program is patented under US patents 10,998,087 and 11,114,183.

For material identification, characterization, and quantification, it is useful to estimate system-independent material properties that do not depend on the detailed specifications of the X-ray computed tomography (CT) system such as spectral response. System independent ρ_e and Z_e (SIRZ) refers to a suite of methods for estimating the system independent material properties of electron density (ρ_e) and effective atomic number (Z_e) of an object scanned using dual-energy X-ray CT (DECT). The current state-of-the-art approach, SIRZ-2, makes certain approximations that lead to inaccurate estimates for large atomic numbered (Z_e) materials. In this paper, we present an extension, SIRZ-3, which iteratively reconstructs the unknown ρ_e and Z_e while avoiding the limiting approximations made by SIRZ-2. Unlike SIRZ-2, this allows SIRZ-3 to accurately reconstruct ρe and Z_e even at large Z_e. SIRZ-3 relies on the use of a new non-linear differentiable forward measurement model that expresses the DECT measurement data as a direct analytical function of ρ_e and Z_e. Leveraging this new forward model, we use an iterative optimization algorithm to reconstruct (or solve for) ρ_e and Z_e directly from the DECT data. Compared to SIRZ-2, we show that the magnitude of performance improvement using SIRZ-3 increases with increasing values for Z_e.

In this study, an efficient end-to-end material classification is proposed for dual energy x-ray imaging devices. Performing prompt geometric and radiometric calibrations, we exploit polynomial modeling on low-high energy ratios to estimate effective atomic numbers (EAN) of the objects, that is based and experimented over twentyfive different materials. Special attention is devoted for dense materials on which the ratio polynomial modeling performs poorly as the thickness increases. A novel material peeling approach is also proposed that uncovers blocked or encapsulated objects and enable precise EAN estimation in cluttered images. The proposed approach provides visually informative x-ray image segmentation.

Understanding the material composition everywhere in a three-dimensional volume is important for medical, security, and material science applications. Using a fan beam geometry with detector-side coded aperture, we demonstrate fast, high-resolution 3D X-ray diffraction (XRD) imaging. The XRD imaging system has a 15 x 15 cm² field of view with a spatial resolution of approximately 1x1.5x7 mm³ (width x length x depth), a fractional momentum transfer resolution of approximately 10%, and scan times on the order of 10 minutes. Using this system, we show the ability to differentiate between two similar-density organic materials (water and PLA) in 3D using conventional, off-the-shelf components.

We propose a framework and impact of applying Machine Learning-based generated imagery to augment data variations for firearm detection in cargo x-ray images. Deep learning-based approaches for object detection have rapidly become the state-of-art and crucial technology for non-intrusive inspection (NII) based on x-ray radiography. The technology is widely employed to reduce or replace tedious labor-intensive inspection to verify cargo content and intercept potential threats at border crossings, ports, and other critical infrastructure facilities. However, the need for variations in the threat cargo content makes accumulating training data for such a system an increasing development cost. Even though threat image projection (TIP) is widely employed to simplify the process into artificially projecting the known threat, a considerable amount of threat object appearances is still needed. To further reduce the cost, we explore the use of GenerativeAdversarial-Network (GAN) to aid dataset creation. GAN is a successful deep learning technique for generating photo-real imagery in many domains. We propose a three-stage training framework dedicated to firearm detection. First, GAN is trained to generate variations of X-ray firearm appearance from binary masks for better image quality compared to the commonly used random noise. Second, the detection training dataset is created in combinations of generated images and actual firearms using TIP. Finally, the dataset is used to train RetinaNet for the detection. Our evaluations reveal that GAN can reduce the training cost in increase detection performance as using the combination of the real and generated firearms increase performance for unseen firearms detection.

The growing x-ray detection burden for vehicles at Ports of Entry in the US requires the development of efficient and reliable algorithms to assist human operator in detecting contraband. Developing algorithms for large-scale non-intrusive inspection (NII) that both meet operational performance requirements and are extensible for use in an evolving environment requires large volumes and varieties of training data, yet collecting and labeling data for these enivornments is prohibitively costly and time consuming. Given these, generating synthetic data to augment algorithm training has been a focus of recent research. Here we discuss the use of synthetic imagery in an object detection framework, and describe a simulation based approach to determining domain-informed threat image projection (TIP) augmentation.

Inspection of commerce entering the country is currently an onerous, custom and extremely labor-intensive process. The imaging of commerce with 2D x-rays aids the ability to interdict certain items entering the country, reducing risk and increasing national security. We use a variety of ML techniques together to develop automated threat recognition (ATR) algorithms to assist this process. The challenge is to develop a system that incorporates several approaches simultaneously and ensemble results to improve overall performance. We employ several algorithmic techniques to solve the problem. Generally, we have synthetic data generation techniques that can rapidly spin up datasets that are then employed to train ML models. GANs are used to refine the synthetic data to resemble reality more faithfully; thresholding objectively improves top line S/N; semi-supervised methods are used to handle data sparsity and leverage any available unlabeled stream of commerce (SOC) data. Suites of object detectors infer in parallel using common techniques and results are then ensembled using a novel graph-based method. Similarity and change detection along with topological data analysis may leverage historical data to investigate anomalies. This paper is an overview of our approach, generating synthetic data and attempting to unite these ML methods to provide the highest performance for ATR of 2D x-ray data.

Photon counting detectors with energy resolving capabilities have the potential to improve computed tomography (CT) imaging and x-ray diffraction (XRD) systems. In order to better understand the use of these detectors in the CT and XRD application spaces, we have experimentally investigated the detector performance of two newly-released photon counting detectors: the Redlen LDA detector and the Kromek D-Matrix v2 detector. Detector performance involves a complicated interplay of semiconductor physics and readout electronics, and the outcome can depend crucially on the properties of the incoming X-rays—specifically the flux and spectral content. Although the LDA and D-Matrix v2 detectors differ in many ways, particularly in the manner in which they collect spectroscopic information, both are of interest for CT and XRD modalities. We report on our analysis of the detector performance, including the noise statistics, detector quantum efficiency, response linearity, and energy resolution of the detectors as well as discuss how our findings influence the use of these detectors in diffraction and transmission measurements.

Dual-energy computed tomography (DECT) is widely used to identify explosives and characterize substances relevant to transportation security screening. The x-ray attenuating behavior of a material can generally be understood in terms of its electron density (ρ_e) and effective atomic number (Z_e). These properties can be estimated from DECT data if the studied material is accompanied by several reference materials of known density and composition. Traditionally, the reference material ρ_e and Z_e values have been used to generate calibration curves with dependencies on the high-energy computed tomography (CT) values and the ratio of low- to high-energy CT values, respectively. However, the accuracy of these 1- D curves, which treat ρe and Z_e as functions of a single variable, breaks down when attempting to characterize materials with high effective atomic number. A more robust calibration technique was developed that instead generates 2-D surfaces from the reference material data. These surfaces treat ρ_e and Z_e as functions of both the high- and low-energy CT values, which better reflects the expected behavior for these properties. Several test materials were packed into bottles and scanned in a commercial explosives detection system along with different sets of reference materials. The results have demonstrated that the developed calibration surfaces generally yield more accurate estimations for test material properties than the traditional approach, particularly for high-Z materials. This work was conducted with the U.S. Department of Homeland Security (DHS) Science and Technology Directorate (S&T) under contract 70RSAT19FR0000016. Any opinions contained herein are those of the author and do not necessarily reflect those of DHS S&T.

Compton backscattering X-rays imaging (CBI) allows the identification of hazardous organic materials by imaging the scattered photons returning from the body after this is illuminated with X-rays. Most of the commercial Compton scanners deployed in airports, base their functionality on a pixel-per-pixel sequential scanning, which is realized by the combination of a fan-beam collimator and a rotating chopper-wheel. This paper explores a simple and faster approach, coined compressive X-ray Compton backscattering imager (CXBI), where the body is illuminated with a static coded cone-beam pattern at once and several snapshots are acquired while the mask, or equivalently the body is translated at constant speed.

X-ray Phase Contrast Imaging (XPCI) is an imaging method that can provide quantitative information about the change in phase of X-ray wavefronts as they pass through an object. XPCI can image objects that cannot be easily seen in conventional absorption imaging, such as thin, weakly-absorbing objects. Most exploration into XPCI has involved synchrotron sources, which are large, fixed facilities and not widely available. Several tabletop methods exist, but these generally rely on interferometric methods or complicated gratings. We began investigating Edge-Illumination (EI), a non-interferometric, inexpensive XPCI method that can use a standard x-ray tube. However, EI requires at least two different spatial shifts, with small aperture openings and precise beam alignment, thereby increasing the complexity of the method. Due to the limitations of EI and the rise in availability of spectrally sensitive detectors, we propose a variant of EI, called Spectrally Responsive Edge Illumination (SREI), which relies on a diversity of X-ray energies instead of spatial shifts. Our goal is to develop an XPCI method that is simple, robust, and easily implementable with commercially available equipment. I will report on our progress.

The Transportation Security Laboratory of the Department of Homeland Security Science and Technology Directorate is tasked with establishing quality control metrics for assessing the performance of x-ray imaging technologies used by airport security. In conformity with that mission, a high-energy x-ray diffraction apparatus was constructed to characterize the xray diffraction properties of common household items as well as explosives and other potential threats. Current x-ray imaging technologies may rely on identifying potential threats in baggage and cargo based on their size, shape, and x-ray attenuating properties. Future technologies are anticipated to include diffraction information that could more specifically identify materials within baggage and cargo based on their chemical and physical composition. As a result, an understanding of the diffraction properties of goods that may be brought onto airplanes is vital for future assessments of these emerging technologies. To gauge the utility of using the diffractometer for quality control evaluations and material identification, a group of common materials were examined, and the dependence of scattering on momentum transfer was measured.

Large scale non-intrusive inspection (NII) of commercial vehicles is being adopted in the U.S. at a pace and scale that will result in a commensurate growth in adjudication burdens at land ports of entry. The use of computer vision and machine learning models to augment human operator capabilities is critical in this sector to ensure the flow of commerce and to maintain efficient and reliable security operations. The development of models for this scale and speed requires novel approaches to object detection and novel adjudication pipelines. Here we propose a notional combination of existing object detection tools using a novel ensembling framework to demonstrate the potential for hierarchical and recursive operations. Further, we explore the combination of object detection with image similarity as an adjacent capability to provide post-hoc oversight to the detection framework. The experiments described herein, while notional and intended for illustrative purposes, demonstrate that the judicious combination of diverse algorithms can result in a resilient workflow for the NII environment.

Chest X-rays can quickly assess the COVID-19 status of test subjects and address the problem of inadequate medical resources in emergency departments and centers. The image classification model established by the deep learning method of artificial intelligence can help doctors make a better judgment on patients with COVID-19 and related lung diseases. We compared and analyzed the current popular deep learning image classification methods, VGGNet, GoogleNet, and ResNet, using publicly available chest X-ray datasets on COVID-19 from different organizations. According to the characteristics of chest X-ray images and the classification results of the deep learning algorithm, a novel image classification algorithm, CovidXNet, is proposed. Based on the ResNet model, the CovidXNet algorithm introduces the hard sample memory pool method to improve the accuracy and generalization of the algorithm. CovidXNet is able to categorize chest X-ray images more efficiently and accurately than other popular image classification algorithms, allowing doctors to quickly confirm the patient’s diagnosis.

X-ray security screening has become crucial to maintaining safety in public spaces. Hence, X-ray screening equipment is widely used in airports, shopping centers, etc. to prevent the transportation of harmful objects. However, these equipment are not capable of detecting threats without human labor. In recent years, automatic threat detection in baggage has been studied and several methods have been offered on X-ray images. In this study, we introduce a publicly-available single view dual-channel X-ray dataset, called the HUMS X-ray dataset, emphasizing the efforts of Hacettepe University and MS Spektral Inc. This dataset includes both the low energy, high energy, as well as the false-colored images of knife threats in baggage under complex scenarios such as occlusion. Then, we detect the threats in both the HUMS dataset and the SIXray datasets using architectures based on Convolutional Neural Network (CNN) techniques. Three popular object detection algorithms, namely the Faster RCNN, YOLOv3 (You Only Look Once), and SSD (Single Shot Detector) are applied on SIXray, the larger X-ray dataset. Then the acquired best model is transferred to the relatively small and different dual X-ray baggage imagery dataset to detect knife threats, using the learned weights from the large X-ray dataset, and the effects of few shot learning and fine tuning is investigated. Furthermore, to observe the effects of the low energy and high energy images, the HUMS X-ray dataset is trained with false-colored, low energy, and high energy images. The dataset is publicly available with all the low energy, high energy and false-colored images.

The Transportation Security Laboratory (TSL) performs testing of explosives detection systems using explosives and other hazardous materials. Inert simulants are also used as substitutes in potentially dangerous testing situations or at testing locations where explosives are prohibited. Each simulant must first be verified that it accurately represents the material on the specific detection platform it was designed for. In addition to the simulant-threat matching, lot-to-lot quality control testing is performed for simulants and threats to ensure that their physical properties remain consistent. Historically, x-ray verification has been limited to using features such as electron density and effective atomic number. While efficient, these features are limited in their application, as they do not provide information related to the material’s structural properties. In this study, four classification methods were tested using imagery-derived texture features to characterize materials and distinguish them from one another. The first three approaches (k-nearest neighbors, support-vector machine, and artificial neural network) were tested using 22 first- and second-order texture features derived from computed tomography images. The fourth method (convolutional neural network) used internally derived features. Based on the test results, a determination was made that the CNN and k-NN were the best algorithms to use to characterize materials based on their texture features.