Imaging Detector Array Research Articles

Adaptive machine learning method for photoacoustic computed tomography based on sparse array sensor data

Background and objectivePhotoacoustic computed tomography (PACT) is a non-invasive biomedical imaging technology that has developed rapidly in recent decades, especially has shown potential for small animal studies and early diagnosis of human diseases. To obtain high-quality images, the photoacoustic imaging system needs a high-element-density detector array. However, in practical applications, due to the cost limitation, manufacturing technology, and the system requirement in miniaturization and robustness, it is challenging to achieve sufficient elements and high-quality reconstructed images, which may even suffer from artifacts. Different from the latest machine learning methods based on removing distortions and artifacts to recover high-quality images, this paper proposes an adaptive machine learning method to firstly predict and complement the photoacoustic sensor channel data from sparse array sampling and then reconstruct images through conventional reconstruction algorithms. MethodsWe develop an adaptive machine learning method to predict and complement the photoacoustic sensor channel data. The model consists of XGBoost and a neural network named SS-net. To handle data sets of different sizes and improve the generalization, a tunable parameter is used to control the weights of XGBoost and SS-net outputs. ResultsThe proposed method achieved superior performance as demonstrated by simulation, phantom experiments, and in vivo experiment results. Compared with linear interpolation, XGBoost, CAE, and U-net, the simulation results show that the SSIM value is increased by 12.83%, 6.78%, 21.46%, and 12.33%. Moreover, the median R2 is increased by 34.4%, 8.1%, 28.6%, and 84.1% with the in vivo data. ConclusionsThis model provides a framework to predict the missed photoacoustic sensor data on a sparse ring-shaped array for PACT imaging and has achieved considerable improvements in reconstructing the objects. Compared with linear interpolation and other deep learning methods qualitatively and quantitatively, our proposed methods can effectively suppress artifacts and improve image quality. The advantage of our methods is that there is no need for preparing a large number of images as the training dataset, and the data for training is directly from the sensors. It has the potential to be applied to a wide range of photoacoustic imaging detector arrays for low-cost and user-friendly clinical applications.

Anti-noise diffractive neural network for constructing an intelligent imaging detector array.

To develop an intelligent imaging detector array, a diffractive neural network with strong robustness based on the Weight-Noise-Injection training is proposed. According to layered diffractive transformation under existing several errors, an accurate and fast object classification can be achieved. The fact that the mapping between the input image and the label in Weight-Noise-Injection training mode can be learned, means that the prediction of the optical network being insensitive to disturbances so as to improve its noise resistance remarkably. By comparing the accuracy under different noise conditions, it is verified that the proposed model can exhibit a higher accuracy.

Characterization of Fine-Pixel X-Ray Imaging Detector Array Fabricated by Using Thick Single-Crystal CdTe Layers on Si Substrates Grown by MOVPE

A novel approach for developing a large area, high-spatial-resolution X-ray imaging detector is presented. This approach uses metal–organic vapor phase epitaxy grown thick single-crystal CdTe layers grown directly on Si substrates. The detector consists of a p-CdTe/n-CdTe/n+-Si heterojunction diode structure, where the n-CdTe and the p-CdTe layers are successively grown on the n+-Si substrate. An array of ( $128\times128$ ) pixels was formed on the p-CdTe side, where each pixel is $60~\mu \text{m}^{\textsf {2}}$ in an 80- $\mu \text{m}$ pixel pitch. The detector array was bump bonded to a charge integration-type CMOS readout ASIC. The basic performance of this detector was evaluated by measuring the dark currents, the X-ray generation currents as well as taking the X-ray images of some objects. The results were promising which confirmed that the detector developed can be applied in high spatial resolution X-ray imaging.

A high-impedance detector-array glove for magnetic resonance imaging of the hand.

Densely packed resonant structures used for magnetic resonance imaging (MRI), such as nuclear magnetic resonance phased-array detectors, suffer from resonant inductive coupling, which restricts coil design to fixed geometries, imposes performance limitations, and narrows the scope of MRI experiments to motionless subjects. Here, we report the design of high-impedance detectors, and the fabrication and performance of a wearable detector array for MRI of the hand, that cloak themselves from electrodynamic interactions with neighboring elements. We experimentally verified that the detectors do not suffer from signal-to-noise degradation mechanisms typically observed with the use of traditional low-impedance elements. The detectors are adaptive and can accommodate movement, providing access to the imaging of soft-tissue biomechanics with unprecedented flexibility. The design of the wearable detector glove exemplifies the potential of high-impedance detectors in enabling a wide range of applications that are not well suited to traditional coil designs.

Random Small Probability Sample Matrix Used in Compressed Sensing Imaging

This investigation aimed to one common key issue of the two imaging types, the sampling of the template configuration. A first selector framework that employs MATLAB to perform attractive features of the small probability matrices in random sampling matrix from Gaussian random matrix was presented. Gaussian random matrix is usually used as random sampling matrix in Single Pixel Camera. The small probability matrices in random sampling matrix can obtain a recovery image of higher accuracy for singe detector imaging system and detector array imaging system.

Computational Study of Integrated Neutron/Photon Imaging for Illicit Material Detection

The feasibility of integration of photon and neutron radiography for nondestructive detection of illicit materials was examined. The MCNP5 code was used to model a radiography system consisting of accelerator-based neutron and photon sources and the imaging detector array, with an object under scrutiny placed between them. For this examination, the objects consisted of a matrix of low-Z and high-Z materials of various shapes and density. Transmission-radiography computations were carried out using 2.5-MeV deuterium-deuterium and 14-MeV deuterium-tritium neutron sources, and a 0.3-MeV photon source. The radiography tallies for both neutron and photon sources were modeled for the same geometry of the system. The photon-to- neutron transmission ratios were determined for each pixel of the detector array and utilized to identify the presence of specific materials in the radiographic images. By focusing on the inherent difference between neutron and photon interactions, it was possible to determine the shape and material composition of complex objects present within a pallet or a shipping container. The use of a single imaging array of scintillation detectors for simultaneous measurements of fast neutrons and photons is discussed, and its function in the dual neutron/photon radiography applications is addressed.

Study of a New Design of P-N Semiconductor Detector Array for Nuclear Medicine Imaging by Monte Carlo Simulation Codes

Gamma camera is an important apparatus in nuclear medicine imaging. Its detection part is consists of a scintillation detector with a heavy collimator. Substitution of semiconductor detectors instead of scintillator in these cameras has been effectively studied. In this study, it is aimed to introduce a new design of P-N semiconductor detector array for nuclear medicine imaging. A P-N semiconductor detector composed of N-SnO2 :F, and P-NiO:Li, has been introduced through simulating with MCNPX monte carlo codes. Its sensitivity with different factors such as thickness, dimension, and direction of emission photons were investigated. It is then used to configure a new design of an array in one-dimension and study its spatial resolution for nuclear medicine imaging. One-dimension array with 39 detectors was simulated to measure a predefined linear distribution of Tc(99_m) activity and its spatial resolution. The activity distribution was calculated from detector responses through mathematical linear optimization using LINPROG code on MATLAB software. Three different configurations of one-dimension detector array, horizontal, vertical one sided, and vertical double-sided were simulated. In all of these configurations, the energy windows of the photopeak were ± 1%. The results show that the detector response increases with an increase of dimension and thickness of the detector with the highest sensitivity for emission photons 15-30° above the surface. Horizontal configuration array of detectors is not suitable for imaging of line activity sources. The measured activity distribution with vertical configuration array, double-side detectors, has no similarity with emission sources and hence is not suitable for imaging purposes. Measured activity distribution using vertical configuration array, single side detectors has a good similarity with sources. Therefore, it could be introduced as a suitable configuration for nuclear medicine imaging. It has been shown that using semiconductor P-N detectors such as P-NiO:Li, N-SnO2 :F for gamma detection could be possibly applicable for design of a one dimension array configuration with suitable spatial resolution of 2.7 mm for nuclear medicine imaging.

Device localization and dynamic scan plane selection using a wireless magnetic resonance imaging detector array

A prototype wireless guidance device using single sideband amplitude modulation (SSB) is presented for a 1.5T magnetic resonance imaging system. The device contained three fiducial markers each mounted to an independent receiver coil equipped with wireless SSB technology. Acquiring orthogonal projections of these markers determined the position and orientation of the device, which was used to define the scan plane for a subsequent image acquisition. Device localization and scan plane update required approximately 30 ms, so it could be interleaved with high temporal resolution imaging. Since the wireless device is used for localization and does not require full imaging capability, the design of the SSB wireless system was simplified by allowing an asynchronous clock between the transmitter and receiver. When coupled to a high readout bandwidth, the error caused by the lack of a shared frequency reference was quantified to be less than one pixel (0.78 mm) in the projection acquisitions. Image guidance with the prototype was demonstrated with a phantom where a needle was successfully guided to a target and contrast was delivered. The feasibility of active tracking with a wireless detector array is demonstrated. Wireless arrays could be incorporated into devices to assist in image-guided procedures.

Three‐dimensional optoacoustic tomography using a conventional ultrasound linear detector array: Whole‐body tomographic system for small animals

Optoacoustic imaging relies on the detection of ultrasonic waves induced by laser pulse excitations to map optical absorption in biological tissue. A tomographic geometry employing a conventional ultrasound linear detector array for volumetric optoacoustic imaging is reported. The geometry is based on a translate-rotate scanning motion of the detector array, and capitalizes on the geometrical characteristics of the transducer assembly to provide a large solid angular detection aperture. A system for three-dimensional whole-body optoacoustic tomography of small animals is implemented. The detection geometry was tested using a 128-element linear array (5.0∕7.0 MHz, Acuson L7, Siemens), moved by steps with a rotation∕translation stage assembly. Translation and rotation range of 13.5 mm and 180°, respectively, were implemented. Optoacoustic emissions were induced in tissue-mimicking phantoms and ex vivo mice using a pulsed laser operating in the near-IR spectral range at 760 nm. Volumetric images were formed using a filtered backprojection algorithm. The resolution of the optoacoustic tomography system was measured to be better than 130 μm in-plane and 330 μm in elevation (full width half maximum), and to be homogenous along a 15 mm diameter cross section due to the translate-rotate scanning geometry. Whole-body volumetric optoacoustic images of mice were performed ex vivo, and imaged organs and blood vessels through the intact abdominal and head regions were correlated to the mouse anatomy. Overall, the feasibility of three-dimensional and high-resolution whole-body optoacoustic imaging of small animal using a conventional linear array was demonstrated. Furthermore, the scanning geometry may be used for other linear arrays and is therefore expected to be of great interest for optoacoustic tomography at macroscopic and mesoscopic scale. Specifically, conventional detector arrays with higher central frequencies may be investigated.

Fabrication of radiation imaging detector arrays using MOVPE grown thick single crystal CdTe layers on Si substrate

AbstractWe present fabrication process of two‐dimensional (2D) energy discriminating pixel‐type imaging arrays. Each pixel in the array possesses a p‐CdTe/n‐CdTe/n+‐Si heterojunction diode structure. The array is fabricated using metalorganic vapour‐phase epitaxy grown thick layers of single crystal CdTe on (211)Si substrates. In order to avoid the charge sharing problem among the adjacent pixels, we used a mechanical saw based pixel pattering technique, which produced pixels that are isolated from each other by narrow vertical trenches. Preliminary evaluation with room‐temperature reverse biased dark current values exhibited uniform response from the entire pixels in the array. This opens a new possibility for developing large‐area imaging arrays in a simple and efficient way. (© 2012 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

Dual-Color InAs/GaSb Superlattice Focal-Plane Array Technology

Within a very few years, InAs/GaSb superlattice technology has proven its suitability for high-performance infrared imaging detector arrays. At the Fraunhofer Institute for Applied Solid State Physics (IAF) and AIM Infrarot-Module GmbH, efforts have been focused on developing mature fabrication technology for dual-color InAs/GaSb superlattice focal-plane arrays for simultaneous, colocated detection at 3 μm to 4 μm and 4 μm to 5 μm in the mid-wavelength infrared atmospheric transmission window. Integrated into a wide-field-of-view missile approach warning system for an airborne platform, a very low number of pixel outages and cluster defects is mandatory for bispectral detector arrays. Process refinements, intense root-cause analysis, and specific test methodologies employed at various stages during the process have proven to be the key for yield enhancements.

Solid-state photon-counting hybrid detector array for high-resolution multi-energy X-ray imaging

We present in this article the development of a photon-counting, energy-discriminating modular detector based on a pixelated CdZnTe sensor coupled pixel-by-pixel to a novel Digital Pixel Sensor (DPS) readout. The detector is designed for munitions inspection, breast X-ray CT and SPECT/MRI. The current DPS design can also be used to read out other solid-state sensors. The prototype detector is 5.5 mm×5.5 mm in size, and consists of 19×19 pixels on a 250 μm pitch. The DPS is designed in a 0.35 μm process, and every pixel includes a preamplifier , a leakage-current subtraction circuit, an auto-zeroed programmable-gain stage, five comparators, a variable-delay reset circuit and five 16 bit counters. The module is expected to operate at high X-ray fluence exceeding 80 MHz/mm 2 , and to improve resolution and contrast in images, while significantly enhancing their signal-to-noise ratio, and assist in identifying material composition via dual-energy imaging. The detector design, fabrication and anticipated performance are discussed.

Feasibility of Localized Substrate Thinning for Reduced Dislocation Density in CdTe/Si Heterostructures

HgCdTe heteroepitaxy on low-cost, large-lattice-mismatched substrates such as Si continue to be plagued by large threading dislocation densities that ultimately reduce the operability of the thermal imaging detector array. Molecular-beam epitaxy (MBE) of 10 μm- to 15 μm-thick CdTe buffer layers has played a crucial role in reducing dislocation densities to current state-of-the-art levels. Herein, we examine the possibility that growth on locally back-thinned substrates could prove advantageous in further reducing dislocation densities in the CdTe/Si heteroepitaxial system. Using defect decoration techniques, a decrease in dislocation (etch-pit) density of up to ~42% has been measured in CdTe regions where the underlying Si substrate was chemically back-thinned to ~20 μm. A theoretical understanding is proposed, where a substrate-thickness-dependent dislocation image force is a likely cause for the experimentally observed reduction in threading dislocation density. These observations raise the prospect of combining localized substrate thinning with other techniques to further reduce dislocation densities to levels sought for HgCdTe/CdTe/Si and other large-lattice-mismatched systems.

STARDUST: A Code for the Simulation of Particle Tracks on Arrays of Sensitive Volumes With Substrate Diffusion Currents

This paper describes STARDUST, a new Monte Carlo code dedicated to the simulation of particle tracks on imaging detector arrays. The geometrical detector model, the particle parameters sampling method, the particle tracking and energy deposition and, finally, the carrier collection by diffusion currents in the substrate are described in detail. Some examples of simulations are given at the end, together with a comparison between COROT satellite data and simulation results.

A thermopile detector array with scaled TE elements for use in an integrated IR microspectrometer

The design and fabrication of a thermopile detector array for use in a fully integrated infrared optical spectrometer are described. IC-compatible MEMS technologies are used for fabrication of the spectrometer components, such as the slit, planar imaging diffraction grating and detector array. The IR micro-spectrometer was designed for operation in the 1.5–3 µm wavelength range with the size of the largest dimension about 8 mm. The imaging properties of the diffraction grating result in non-uniform dispersion, which imposes special requirements on the dimensions of each single detector in the array. The result is an array of unequally sized elements. The design considers technological constraints, sensitivity and cross-talk between elements. Simulation results, final design, fabrication technique and fabricated devices are presented.

Monolithically integrated near-infrared and mid-infrared detector array for spectral imaging

A multi-band focal plane array sensitive in near-infrared (near-IR) and mid-wavelength infrared (MWIR) is been developed by monolithically integrating a near-infrared (1–1.5μm) p–i–n photodiode with a mid-infrared (3–5μm) QWIP. This multiband detector involves both intersubband and interband transitions in III–V semiconductor layer structures. Each detector stack absorbs photons within the specified wavelength band, while allowing the transmission of photons in other spectral bands, thus efficiently permitting multiband detection. Monolithically grown material characterization data and individual detector test results ensure the high quality of material suitable for near-infrared/QWIP dual-band focal plane array.

The X-ray emission from Young Stellar Objects in the ρ Ophiuchi cloud core as seen by XMM-Newton

We observed the main core F of the rho Ophiuchi cloud, an active star-forming region located at ~140 pc, using XMM-Newton with an exposure of 33 ks. We detect 87 X-ray sources within the 30' diameter field-of-view of the it EPIC imaging detector array. We cross-correlate the positions of XMM-Newton X-ray sources with previous X-ray and infrared (IR) catalogs: 25 previously unknown X-ray sources are found from our observation; 43 X-ray sources are detected by both XMM-Newton and Chandra; 68 XMM-Newton X-ray sources have 2MASS near-IR counterparts. We show that XMM-Newton and Chandra have comparable sensitivity for point source detection when the exposure time is set to ~30 ks for both. We detect X-ray emission from 7 Class I sources, 26 Class II sources, and 17 Class III sources. The X-ray detection rate of Class I sources is very high (64 %), which is consistent with previous Chandra observations in this area. We propose that 15 X-ray sources are new class III candidates, which doubles the number of known Class III sources, and helps to complete the census of YSOs in this area. We also detect X-ray emission from two young bona fide brown dwarfs, GY310 and GY141, out of three known in the field of view. GY141 appears brighter by nearly two orders of magnitude than in the Chandra observation. We extract X-ray light curves and spectra from these YSOs, and find some of them showed weak X-ray flares. We observed an X-ray flare from the bona fide brown dwarf GY310. We find as in the previous Chandra observation of this region that Class I sources tend to have higher temperatures and heavier X-ray absorptions than Class II and III sources.

The X-ray emission from Young Stellar Objects in theρOphiuchi cloud core as seen by XMM-Newton

We observed the main core F of the rho Ophiuchi cloud, an active star-forming region located at ~140 pc, using XMM-Newton with an exposure of 33 ks. We detect 87 X-ray sources within the 30' diameter field-of-view of the it EPIC imaging detector array. We cross-correlate the positions of XMM-Newton X-ray sources with previous X-ray and infrared (IR) catalogs: 25 previously unknown X-ray sources are found from our observation; 43 X-ray sources are detected by both XMM-Newton and Chandra; 68 XMM-Newton X-ray sources have 2MASS near-IR counterparts. We show that XMM-Newton and Chandra have comparable sensitivity for point source detection when the exposure time is set to ~30 ks for both. We detect X-ray emission from 7 Class I sources, 26 Class II sources, and 17 Class III sources. The X-ray detection rate of Class I sources is very high (64 %), which is consistent with previous Chandra observations in this area. We propose that 15 X-ray sources are new class III candidates, which doubles the number of known Class III sources, and helps to complete the census of YSOs in this area. We also detect X-ray emission from two young bona fide brown dwarfs, GY310 and GY141, out of three known in the field of view. GY141 appears brighter by nearly two orders of magnitude than in the Chandra observation. We extract X-ray light curves and spectra from these YSOs, and find some of them showed weak X-ray flares. We observed an X-ray flare from the bona fide brown dwarf GY310. We find as in the previous Chandra observation of this region that Class I sources tend to have higher temperatures and heavier X-ray absorptions than Class II and III sources.

TES microcalorimeter readout via transformer

Readout configuration for Transition Edge Sensor (TES) electrically coupled to a low noise warm front-end via transformer is studied. The study was aimed at the implementation of the readout involved in imaging with microcalorimeter detector arrays using frequency multiplexing technique (Appl. Phys. Lett. 81(1) (2002) 159). A model describing both TES electrothermal reaction and the readout response is investigated. Contribution of electronic noise to TES energy resolution is calculated. Prospective readout architecture with TES electrical biasing from the warm readout side is introduced. The architecture allows building of large imaging detector arrays with TES. It is shown that an unprecedented combination of imaging and spectrometry features can be achieved with TES readout via transformer.

A low-background [formula omitted] bolometer array test facility and its use in evaluating prototype arrays for FIRST-SPIRE

We report on the imaging detector array test and evaluation program for the Spectral and Photometric Imaging REceiver (SPIRE) instrument for the Far InfraRed and Submillimetre Telescope (FIRST) satellite. Three candidate detector array technologies are under evaluation: (i) hexagonally close-packed arrays of 2 Fλ feedhorns with NTD germanium sensors; (ii) filled absorber arrays with 0.5 Fλ square pixels with ion-implanted silicon sensors; and (iii) 0.5 Fλ filled arrays with transition-edge superconducting (TES) sensors. Detector arrays are required for SPIRE bands covering the 200–700 μm range. To meet the SPIRE performance requirements, the detectors must be background limited by the thermal emission from the 80 K low-emissivity FIRST telescope, achieve a speed of response of 20 Hz, and be fabricated in sizes up to 32×32. A novel cryogenic array test facility has been built to enable a range of experiments to be carried out under low background to enable a comprehensive evaluation and calibration programme on prototype detector arrays with a minimum of disruptive warm-up/cool-down cycles. The facility will be used to compare the performance of the different array types prior to array technology selection for SPIRE.