Radioisotope Identification Research Articles

FPGA-based fast bin-ratio spiking ensemble network for radioisotope identification

In this work, we demonstrate the training, conversion, and implementation flow of an FPGA-based bin-ratio ensemble spiking neural network applied for radioisotope identification. The combination of techniques including learned step quantisation (LSQ) and pruning facilitated the implementation by compressing the network’s parameters down to 30% yet retaining the accuracy of 97.04% with an accuracy loss of less than 1%. Meanwhile, the proposed ensemble network of 20 3-layer spiking neural networks (SNNs), which incorporates 1160 spiking neurons, only needs 334 μs for a single inference with the given clock frequency of 100 MHz. Under such optimisation, this FPGA implementation in an Artix-7 board consumes 157 μJ per inference by estimation.

Nuclide identification algorithm for Polyvinyl Toluene scintillation detector based on Deep Neural Network

Radiation portal monitors (RPMs) are now stationed at strategic areas (airports, ports, etc.) to identify the illegal transportation of radioactive sources and nuclear items. RPMs are typically fitted with a PVT detector with a high recording efficiency. Radioisotope identification from the gamma spectrum acquired on this detector is normally not regarded due to the low resolution. This research describes an artificial neural network-based isotope identification algorithm that was applied to the gamma spectrum collected from the RPM's PVT detector. With excellent precision, this approach can detect one or a mixture of isotopes on the spectrum. The model still recognizes the training isotopes with >89 percent accuracy for spectra with the gain displacement in the range of 20 percent.

Application of peak based-Bayesian statistical method for isotope identification and categorization of depleted, natural and low enriched uranium measured by LaBr3:Ce scintillation detector

Todays, medium energy resolution detectors are preferably used in radioisotope identification devices(RID) in nuclear and radioactive material categorization. However, there is still a need to develop or enhance « automated identifiers » for the useful RID algorithms. To decide whether any material is SNM or NORM, a key parameter is the better energy resolution of the detector. Although masking, shielding and gain shift/stabilization and other affecting parameters on site are also important for successful operations, the suitability of the RID algorithm is also a critical point to enhance the identification reliability while extracting the features from the spectral analysis. In this study, a RID algorithm based on Bayesian statistical method has been modified for medium energy resolution detectors and applied to the uranium gamma-ray spectra taken by a LaBr3:Ce detector. The present Bayesian RID algorithm covers up to 2000 keV energy range. It uses the peak centroids, the peak areas from the measured gamma-ray spectra. The extraction features are derived from the peak-based Bayesian classifiers to estimate a posterior probability for each isotope in the ANSI library. The program operations were tested under a MATLAB platform.The present peak based Bayesian RID algorithm was validated by using single isotopes(241Am, 57Co, 137Cs, 54Mn, 60Co), and then applied to five standard nuclear materials(0.32–4.51% at.235U), as well as natural U- and Th-ores. The ID performance of the RID algorithm was quantified in terms of F-score for each isotope. The posterior probability is calculated to be 54.5–74.4% for 238U and 4.7–10.5% for 235U in EC-NRM171 uranium materials. For the case of the more complex gamma-ray spectra from CRMs, the total scoring (ST) method was preferred for its ID performance evaluation. It was shown that the present peak based Bayesian RID algorithm can be applied to identify 235U and 238U isotopes in LEU or natural U–Th samples if a medium energy resolution detector is was in the measurements.

核検知及び核セキュリティ事案初動対応を支援する深層ニューラルネットワークを用いた核種判定アルゴリズム

Rapid and precise radioisotope identification in the scene of nuclear detection and nuclear security incidents is one of the challenging issues for the prompt response to the detection alarm or the incidents. A radioisotope identification algorithm using a deep artificial neural network model applicable to handheld γ-ray detectors has been proposed in the present paper. The proposed algorithm automatically identifies gamma-emitting radioisotopes based on the count contribution ratio (CCR) from each of them estimated by the deep artificial neural network model trained by simulated γ-ray spectra. The automated radioisotope identification algorithm can support first responders of nuclear detection and nuclear security incidents without sufficient experience and knowledge in radiation measurement. The authors tested the performance of the proposed algorithm using two different types of deep artificial neural network models in application to handheld detectors having high or low energy resolution. The proposed algorithm showed high performance in identifying artificial radioisotopes for actually measured γ-ray spectra. It was also confirmed that the algorithm is applicable to identifying 235U and automated uranium categorization by analyzing estimated CCRs by the deep artificial neural network models. The authors also compared the performance of the proposed algorithm with a conventional radioisotope identification method and discussed promising ways to improve the performance of the algorithm using the deep artificial neural network.

Assessment of Temporary Radioactivation for Tissue Expanders in Breast Radiation Therapy: Preliminary Study

Background: As breast tissue expanders consist of metallic materials in the needle guard and ferromagnetic injection port, irradiation can produce radioactivation.Materials and Methods: A CPX4 (Mentor Worldwide LLD) breast tissue expander was exposed using the Versa HD (Elekta) linear accelerator. Two photon energies of 6 and 10 MV-flattening filter free (FFF) beams with 5,000 monitor units (MU) were irradiated to identify the types of radiation. Furthermore, 300 MU with 10 MV-FFF beam was exposed to the CPX4 breast tissue expander by varying the machine dose rates (MDRs) 600, 1,200, and 2,200 MU/ min. To assess the instantaneous dose rates (IDRs) solely from the CPX4, a tissue expander was placed outside the treatment room after beam irradiation, and a portable radioisotope identification device was used to identify the types of radiation and measure IDR.Results and Discussion: After 5,000 MU delivery to the CPX4 breast tissue expander, the energy spectrum whose peak energy of 511 keV was found with 10 MV-FFF, while there was no resultant one with 6 MV-FFF. The time of each measurement was 1 minute, and the mean IDRs from the 10 MV-FFF were 0.407, 0.231, and 0.180 μSv/hr for the three successive measurements. Following 10 MV-FFF beam irradiation with 300 MU indicated around the background level from the first measurement regardless of MDRs.Conclusion: As each institute room entry time protocol varies according to the working hours and occupational doses, we suggest an addition of 1 minute from the institutes’ own room entry time protocol in patients with CPX4 tissue expander and the case of radiotherapy vaults equipped with a maximum energy of 10 MV photon beams.

ANALYSIS OF THE RADIOISOTOPES RECOGNITION POSSIBILITY BY MEANS OF THE ABSORBED DOSE MEASUREMENT WITH DOSIMETRIC DETECTORS OF DIFFERENT DENSITY

The work is devoted to the problem of identifying an unknown source of g-radiation as a task of emergency dosimetry in the case of the use for terrorist purposes of the radiation-dispersive devices, known also as dirty bomb. The possibility of identifying an unknown g-source is considered based on the energy dependence of the absorption of ionizing photon radiation, which in passive dosimetry manifests itself as dosimetric sensitivity. Radioisotope recognition is analyzed by the ratio of dosimetric sensitivities of heavy (with a high value of the effective atomic number Zeff) and light (with a low Zeff value) dosimetric detectors, for the values of g-radiation energies of those radioisotopes that can probably be used to create a dirty bomb. As a light dosimetric detector, BeO ceramics is considered, and as a heavy one, a dosimetric detector based on single crystalline yttrium-aluminum perovskite (YAlO3) or lutetium-aluminum perovskite (LuAlO3). The influence of the accuracy of the absorbed dose measurement on the reliability of radioisotope identification is discussed and approaches for its practical implementation are proposed.

Pulse height estimation and pulse shape discrimination in pile-up neutron and gamma ray signals from an organic scintillation detector using multi-task learning

We developed a multi-tasking deep learning model for simultaneous pulse height estimation and pulse shape discrimination for pile-up n/γ signals. Compared with single-tasking models, our model showed better spectral correction performance with higher recall for neutrons. Further, it achieved more stable neutron counting with less signal loss and a lower error rate in the predicted gamma ray spectra. Our model can be applied to a dual radiation scintillation detector to discriminatively reconstruct each radiation spectrum for radioisotope identification and quantitative analysis.

Identification of multiple radioisotopes through convolutional neural networks trained on 2-D transformed gamma spectral data from CsI(Tl) spectrometer

A study is conducted to improve the performance of Convolutional Neural Networks (CNNs)-based radioisotope identification (RIID). This study introduces a novel method to transform the dimensional structure of a radioisotope (RI) mixture spectra as training inputs, for the purpose of extending the receptive field and improving RIID performance of the model. The one-dimensional (1-D) composite spectra of radioisotopic mixtures as a dataset to be used as train and test data for a CNNs model are generated by linearly combining the normalized element gamma-ray spectra of single RIs obtained experimentally with the developed CsI(Tl) spectrometer. One more dataset for training and testing another a CNNs model is produced by reshaping the same 1-D composite spectra into two-dimensional (2-D) image data with multiple columns and rows. After the structure and hyper-parameters of the model are determined, each model is trained and tested with each of the two prepared datasets to simultaneously predict the probabilities as the relative count distributions for eight RI classes. The RIID performance of the CNNs model trained with the reshaped 2-D data is compared with that of the model trained with the 1-D data. The performance of each trained model is evaluated by mean magnitude of relative error (MMRE) as performance evaluation metrics. The results show that the model trained with 2-D inputs outperforms the model trained with 1-D inputs and effectively identifies unknown multiple radioisotopes. The proposed method to transform 1-D data into 2-D data represents a promising approach for the CNNs-based RIID of a scintillator spectrometer, especially when multiple peaks or RIs need to be resolved.

A Low-Power Radiation Detection SoC With Neural Network Accelerator for Radioisotope Identification

This work presents a low-power radiation detection and radioisotope identification System on Chip (SoC) featuring mixed-signal sensory electronics and on-chip neural network (NN) acceleration hardware. The chip electronics form a multichannel analyzer (MCA) to process the signal from an external radiation detector and produce an energy spectrum of the incoming radiation detections. The NN hardware accelerator then provides a novel means of using the spectrum to perform radioisotope identification. The chip is, thus, able to determine which radioisotopes are present in the radiation source. An on-chip 32-bit microcontroller (MCU) configures the analog front-end (AFE) electronics, oversees the NN hardware, and provides an interface to retrieve the histogram and NN data. The chip is fabricated on a 65-nm CMOS technology and consumes 8.8 mW while acquiring a histogram. In the tested configuration, the radioisotope identification NN takes <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$694~\mu \text{s}$ </tex-math></inline-formula> to execute, consuming <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$1.94~\mu \text{J}$ </tex-math></inline-formula> of energy and using 13.8 kB of RAM in the process. After training, the system correctly identified which isotopes were present in all histograms in a test set composed of combinations of four trained isotopes.

Automatic energy calibration of radiation portal monitors using deep learning and spectral remapping

Radiation portal monitors comprising large-volume plastic scintillators are commonly used to monitor the smuggling of radioactive materials. Various applications have been proposed to perform radioisotope identification using these monitors. Such applications require calibration of the spectrum measured by the detector to obtain the physical energy spectrum. The relationship between the multichannel readout and energy bins depends on environmental conditions: it implies that energy calibration in radiation portal monitors should be performed periodically, even multiple times in a single day, thus demanding for a simple and fast energy calibration method. In this study, a deep learning model and a spectral remapping method were used to transform the raw detector output into an energy spectrum with constant energy bins. The deep learning model was designed to predict energy calibration parameters based on the channel spectrum of a single radioisotope. The dataset used to train the deep learning model was generated using the spectrum of the radiation portal monitor. A convolutional neural network model was utilized to evaluate the performance. The remapping method was designed to remap calibrated energy bins to fixed energy bins based on the linear interpolation of nearby bins. The performance of the neural network model and of the remapping method were then evaluated based on several measured spectra taken with different conditions, and found to be adequate to fulfill the requirements.

Explanation of Deep Learning–Based Radioisotope Identifier for Plastic Scintillation Detector

Radioisotope identification (RIID) is a representative application of deep learning for radiation measurements. Deep learning-based RIID models have been implemented in various types of radiation detectors; however, very few of these models have been interpreted using explainable artificial intelligence (XAI) methods. This paper presents an explanation of a deep learning–based RIID model for a plastic scintillation detector. The RIID task is defined as a multilabel binary classification problem, and the dataset is generated using a random sampling procedure. The identification performance is verified using experimental data. The experimental results demonstrate that the performance of the RIID models increased with the increase in the total counts of the dataset. Additionally, XAI methods are implemented, and their explanatory performance is verified for the spectral input. The domain knowledge of RIID for the plastic scintillation detector is that patterns near the Compton edge can be used as evidence for the existence of radioisotopes. Among the implemented XAI methods, integrated gradient and layerwise relevance propagation exhibited concurrence with the domain knowledge, with the Shapley value explanation method presenting the most reliable results.

Path-Planning System for Radioisotope Identification Devices Using 4π Gamma Imaging Based on Random Forest Analysis

We developed a path-planning system for radiation source identification devices using 4π gamma imaging. The estimated source location and activity were calculated by an integrated simulation model by using 4π gamma images at multiple measurement positions. Using these calculated values, a prediction model to estimate the probability of identification at the next measurement position was created by via random forest analysis. The path-planning system based on the prediction model was verified by integrated simulation and experiment for a 137Cs point source. The results showed that 137Cs point sources were identified using the few measurement positions suggested by the path-planning system.

Comparison of machine learning approaches for radioisotope identification using NaI(TI) gamma-ray spectrum

This research aims at comparing the performance of different machine learning algorithms used for NaI(TI) gamma-ray detector based radioisotope identification. Six machine learning algorithms were implemented, including support vector machine (SVM), k-nearest neighbor (KNN), logistic regression (LR), naive Bayes (NB), decision tree (DT), and multilayer perceptron (MLP). The hyper-parameters of each model were elaborately optimized. The effects of data size, statistical fluctuation, and spectrum drift were considered. Results show that for smaller data size (5 types of radioisotopes and 6000 spectra), the support vector machine and the logistic regression classifier exhibit higher identification accuracy with less training/predicting time. Whereas for larger data size (14 types of radioisotopes corresponding to the standard IEC 62327–2017), the multilayer perceptron showed highest accuracy but required the longest time for model training. The naive Bayes classifier and the decision tree were prone to make mistakes when fluctuations and distortions were added to the spectra. The k-nearest neighbor classifier, though showing high accuracy for most data sets, consumed the longest time while making prediction.

Comparison of Machine Learning-Based Radioisotope Identifiers for Plastic Scintillation Detector

Background: Identification of radioisotopes for plastic scintillation detectors is challenging because their spectra have poor energy resolutions and lack photo peaks. To overcome this weakness, many researchers have conducted radioisotope identification studies using machine learning algorithms; however, the effect of data normalization on radioisotope identification has not been addressed yet. Furthermore, studies on machine learning-based radioisotope identifiers for plastic scintillation detectors are limited.Materials and Methods: In this study, machine learning-based radioisotope identifiers were implemented, and their performances according to data normalization methods were compared. Eight classes of radioisotopes consisting of combinations of 22Na, 60Co, and 137Cs, and the background, were defined. The training set was generated by the random sampling technique based on probabilistic density functions acquired by experiments and simulations, and test set was acquired by experiments. Support vector machine (SVM), artificial neural network (ANN), and convolutional neural network (CNN) were implemented as radioisotope identifiers with six data normalization methods, and trained using the generated training set.Results and Discussion: The implemented identifiers were evaluated by test sets acquired by experiments with and without gain shifts to confirm the robustness of the identifiers against the gain shift effect. Among the three machine learning-based radioisotope identifiers, prediction accuracy followed the order SVM &gt;ANN&gt;CNN, while the training time followed the order SVM&gt;ANN&gt;CNN.Conclusion: The prediction accuracy for the combined test sets was highest with the SVM. The CNN exhibited a minimum variation in prediction accuracy for each class, even though it had the lowest prediction accuracy for the combined test sets among three identifiers. The SVM exhibited the highest prediction accuracy for the combined test sets, and its training time was the shortest among three identifiers.

Radioisotope identification using sparse representation with dictionary learning approach for an environmental radiation monitoring system

A radioactive isotope identification algorithm is a prerequisite for a low-resolution scintillation detector applied to an unmanned radiation monitoring system. In this paper, a sparse representation with dictionary learning approach is proposed and applied to plastic gamma-ray spectra. Label-consistent K-SVD was used to learn a discriminative dictionary for the spectra corresponding to a mixture of four isotopes (133Ba, 22Na, 137Cs, and 60Co). A Monte Carlo simulation was employed to produce the simulated data as learning samples. Experimental measurement was conducted to obtain practical spectra. After determining the hyper parameters, two dictionaries tailored to the learning samples were tested by varying with the source position and the measurement time. They achieved average accuracies of 97.6% and 98.0% for all testing spectra. The average accuracy of each dictionary was above 96% for spectra measured over 2 s. They also showed acceptable performance when the spectra were artificially shifted. Thus, the proposed method could be useful for identifying radioisotopes in gamma-ray spectra from a plastic scintillation detector even when a dictionary is adapted to only simulated data. Furthermore, owing to the outstanding properties of sparse representation, the proposed approach can easily be built into an in-situ monitoring system.

Convolutional Neural Networks for Challenges in Automated Nuclide Identification.

Improvements in Radio-Isotope IDentification (RIID) algorithms have seen a resurgence in interest with the increased accessibility of machine learning models. Convolutional Neural Network (CNN)-based models have been developed to identify arbitrary mixtures of unstable nuclides from gamma spectra. In service of this, methods for the simulation and pre-processing of training data were also developed. The implementation of 1D multi-class, multi-label CNNs demonstrated good generalisation to real spectra with poor statistics and significant gain shifts. It is also shown that even basic CNN architectures prove reliable for RIID under the challenging conditions of heavy shielding and close source geometries, and may be extended to generalised solutions for pragmatic RIID.

Gamma ray interactions based optimization algorithm: Application in radioisotope identification

This work proposes a new efficient meta-heuristic optimization algorithm called Gamma Ray Interactions Based Optimization (GRIBO). The algorithm mimics different energy loss processes of a gamma-ray photon during its passage through a matter. The proposed novel algorithm has been applied to search for the global minima of 30 standard benchmark functions. The paper also considers solving real optimization problem in the field of nuclear engineering, radioisotope identification. The results are compared with those obtained by the Particle Swarm Optimization, Genetic Algorithm, Gravitational Search Algorithm and Grey Wolf Optimizer algorithms. The comparisons indicate that the GRIBO algorithm is able to provide very competitive results compared to other well-known meta-heuristics.

Self-powered multilayer radioisotope identification device.

This is a computational study to develop a rugged self-powered Radioisotope Identification Device (RIID). The principle of operation relies on the High Energy Current (HEC) concept (Zygmanski and Sajo, Med Phys. 43 4-15, 2016) with measurement of fast electron currents between low-Z and high-Z thin-film electrodes separated by nanoporous aerogel films in a multilayer detector structure whose prototypes were previously investigated (Brivio, Albert, Freund, Gagne, Sajo and Zygmanski, Med Phys, 46 4233-4240, 2019), (Brivio, Albert, Gagne, Freund, Sajo and Zygmanski, J Phys D Appl Phys, 53 265303, 2020). Here, we present an optimal detector design that accounts for a wide energy range (keV-MeV) of x-ray-emitting radioisotopes that are of interest to national security and radiation therapy. We studied numerous multilayer detector geometries with N=1..24 basic detector elements composed of 3 electrodes: N x (Al-aerogel-Ta-aerogel-Al). The thicknesses of electrodes and their total number were varied depending on the incident x-ray spectra and its ability to penetrate and interact with the different layers, producing fast electrons. We used radiation transport simulations to find a balanced geometry that accounts for all energies from 10keV to 6MeV in a single design with relatively few detector elements (N=24). In the balanced design, the electrodes have increasing thickness as a function of depth in the detector, ranging from 0.5μm-Ta and 10μm-Al at the entrance to 10mm-Ta and 2.5mm-Al at the exit. Aerogel thickness was fixed at 50μm. Electron currents forming RIID signals were acquired from all Ta electrodes. A model function M(x, Ei ) representing the detector yield as a function of the cumulative Ta thickness (x) for 70 monoenergetic incident beams (E) was derived. We also investigated the detector response to selected radioactive isotopes (Pd-103, I-125, Pu-239, U-235, Ir-192, Cs-137, Co-60). Additional studies were performed with Bremsstrahlung spectra produced by electron beams in kVp tubes and in MV Linacs used in radiology and radiation therapy departments. We investigated different algorithms for radioisotope identification that would work for unknown unshielded as well as shielded sources. Characteristic features of response functions for monoenergetic beams and radioisotopes were determined and used to develop two inverse algorithms of radioisotope identification. Using these algorithms, we were able to identify the unshielded and shielded sources, quantify the minimum, mean and maximum effective energies of the shielded spectra, and estimate the amount of Compton background in the spectrum. A multilayer sensor based on fast electron current was optimized and studied in its abilities as RIID. A balanced design permits the identification of radioisotopes with of a wide range of keV-MeV energies. The device is low cost, rugged, self-powered and can withstand very high dose rates, allowing deployment in difficult conditions, including radiation incidents. The algorithm we developed for radioisotope identification and spectral unfolding is robust and it is an important component in practical applications.

Generalised gamma spectrometry simulator for problems in nuclide identification

Improvements in Radio-Isotope IDentification (RIID) algorithms have always been a continuous research focus. However, significant developments in machine learning have recently sparked renewed interest. To provide a rapid development environment for this, a generalised gamma simulator has been built using the GEANT4 toolkit. This enables consideration of a diverse range of radiation sources and shielding scenarios. The simulator currently provides training data for the development of neural network based RIID models.

Application of the Approximate Bayesian Computation Algorithm to Gamma-Ray Spectroscopy

Radioisotope identification (RIID) algorithms for gamma-ray spectroscopy aim to infer what isotopes are present and in what amounts in test items. RIID algorithms either use all energy channels in the analysis region or only energy channels in and near identified peaks. Because many RIID algorithms rely on locating peaks and estimating each peak’s net area, peak location and peak area estimation algorithms continue to be developed for gamma-ray spectroscopy. This paper shows that approximate Bayesian computation (ABC) can be effective for peak location and area estimation. Algorithms to locate peaks can be applied to raw or smoothed data, and among several smoothing options, the iterative bias reduction algorithm (IBR) is recommended; the use of IBR with ABC is shown to potentially reduce uncertainty in peak location estimation. Extracted peak locations and areas can then be used as summary statistics in a new ABC-based RIID. ABC allows for easy experimentation with candidate summary statistics such as goodness-of-fit scores and peak areas that are extracted from relatively high dimensional gamma spectra with photopeaks (1024 or more energy channels) consisting of count rates versus energy for a large number of gamma energies.