scispace - formally typeset
Search or ask a question

Showing papers by "Sarah E. Bohndiek published in 2007"


Journal ArticleDOI
TL;DR: The results suggest that CMOS active pixel sensors when coupled to structured CsI:Tl can be used for conventional and advanced digital mammography due to their low noise, high resolution performance.
Abstract: Monolithic complementary metal oxide semiconductor (CMOS) active pixel sensors with high performance have gained attention in the last few years in many scientific and space applications. In order to evaluate the increasing capabilities of this technology, in particular where low dose high resolution x-ray medical imaging is required, critical electro-optical and physical x-ray performance evaluation was determined. The electro-optical performance includes read noise, full well capacity, interacting quantum efficiency, and pixels cross talk. The x-ray performance, including x-ray sensitivity, modulation transfer function, noise power spectrum, and detection quantum efficiency, has been evaluated in the mammographic energy range. The sensor is a 525x525 standard three transistor CMOS active pixel sensor array with more than 75% fill factor and 25x25 {mu}m pixel pitch. Reading at 10 f/s, it is found that the sensor has 114 electrons total additive noise, 10{sup 5} electrons full well capacity with shot noise limited operation, and 34% interacting quantum efficiency at 530 nm. Two different structured CsI:Tl phosphors with thickness 95 and 115 {mu}m, respectively, have been optically coupled via a fiber optic plate to the array resulting in two different system configurations. The sensitivity of the two different system configurations was 43 and 47 electrons permore » x-ray incident on the sensor. The MTF at 10% of the two different system configurations was 9.5 and 9 cycles/mm with detective quantum efficiency of 0.45 and 0.48, respectively, close to zero frequency at {approx}0.44 {mu}C/kg (1.72 mR) detector entrance exposure. The detector was quantum limited at low spatial frequencies and its performance was comparable with high resolution a:Si and charge coupled device based x-ray imagers. The detector also demonstrates almost an order of magnitude lower noise than active matrix flat panel imagers. The results suggest that CMOS active pixel sensors when coupled to structured CsI:Tl can be used for conventional and advanced digital mammography due to their low noise, high resolution performance.« less

59 citations


Journal ArticleDOI
TL;DR: This work aims to reduce sample categorization errors by combining spectral diffraction signature collection with sample imaging, giving more detailed data on the composition of each sample, with a correlation between measured x-ray diffraction images and stained histopathological tissue sections.
Abstract: Identification of specific tissue types in conventional mammographic examinations is extremely limited. However, the use of x-ray diffraction effects during imaging has the potential to characterize the tissue types present due to the fact that each tissue type produces its own unique diffraction signature. Nevertheless, the analysis and categorization of these diffraction signatures by tissue type can be hampered by the inhomogeneous nature of breast tissue, leading to categorization errors where several types are present. This work aims to reduce sample categorization errors by combining spectral diffraction signature collection with sample imaging, giving more detailed data on the composition of each sample. Diffraction microCT was carried out on 19 unfixed breast tissue samples using an energy resolving translate-rotate CT system. High-resolution transmission microCT images were also recorded for comparison and sample composition analysis. Following imaging, the samples were subjected to histopathological analysis. Reconstructing on various momentum transfer regions allows different tissue types to be identified in the diffraction images. Results show a correlation between measured x-ray diffraction images and stained histopathological tissue sections. X-ray diffraction signatures generated from the measured data were categorized and analysed, with a t-test indicating that they have the potential for use in tissue type identification.

43 citations


Journal ArticleDOI
TL;DR: Vanilla could be employed for applications where only a small portion of the image contains relevant information, while OPIC is suited to high-speed imaging applications.
Abstract: A United Kingdom consortium MI3 is founded to develop advanced CMOS image sensors for scientific applications. "Vanilla," a 520520 array of active pixels with 25-m pitch is fabricated in the 0.35- m4 M2P4 metal, 2 poly CMOS process and uses a 3.3-V sup- ply. It has flushed reset circuitry to attain low reset noise and random pixel access for high-speed region-of-interest ROI readout. "OPIC" is a 6472 test structure array of digital pixels with 30-m pitch, fabricated in 0.25- m5 M1P5 metal 1 poly CMOS process with a 3.3/2.5-V sup- ply. It can perform thresholding via an in-pixel comparator for sparse readout at a high frame rate. Characterization of both sensors is per- formed under optical illumination and x-ray exposure. For x-ray charac- terization, both sensors were coupled to a structured thallium-doped ce- sium iodide CsI:Tl scintillator via a fiber optic plate. Vanilla has been found to exhibit 34±3eread noise and a spectral response of 225±5 mA/W at 500 nm and can read a 66 ROI at 24,395 frames/s. OPIC has 46±3eread noise and can perform sparse readout at up to 3700 frames/s. Based on these results, Vanilla could be employed for applications where only a small portion of the image contains relevant information, while OPIC is suited to high-speed imaging applications.

31 citations


Journal ArticleDOI
TL;DR: In this paper, a set of phase-contrast images of different biological samples was acquired by means of the well-established free-space propagation XPCi technique and the obtained results are very encouraging and are in optimum agreement with the predictions of a simulation recently developed by some of the authors thus further supporting its reliability.
Abstract: The aim of the present work is to achieve a first step towards combining the advantages of an innovative X-ray imaging technique—phase-contrast imaging (XPCi)—with those of a new class of sensors, i.e. CMOS-based active pixel sensors (APSs). The advantages of XPCi are well known and include increased image quality and detection of details invisible to conventional techniques, with potential application fields encompassing the medical, biological, industrial and security areas. Vanilla, one of the APSs developed by the MI-3 collaboration (see http://mi3.shef.ac.uk ), was thoroughly characterised and an appropriate scintillator was selected to provide X-ray sensitivity. During this process, a set of phase-contrast images of different biological samples was acquired by means of the well-established free-space propagation XPCi technique. The obtained results are very encouraging and are in optimum agreement with the predictions of a simulation recently developed by some of the authors thus further supporting its reliability. This paper presents these preliminary results in detail and discusses in brief both the background to this work and its future developments.

21 citations


Journal ArticleDOI
TL;DR: In this article, a monolithic active pixel sensor, Vanilla, with 520 × 520 pixels (25 μ m square), has been characterised for the first time Optimisation of the sensor operation was made through variation of frame rates, integration times and on-chip biases and voltages.
Abstract: Novel features of a new monolithic active pixel sensor, Vanilla, with 520 × 520 pixels ( 25 μ m square) has been characterised for the first time Optimisation of the sensor operation was made through variation of frame rates, integration times and on-chip biases and voltages Features such as flushed reset operation, ROI capturing and readout modes have been fully tested Stability measurements were performed to test its suitablility for long-term applications These results suggest the Vanilla sensor—along with bio-medical and space applications—is suitable for use in particle physics experiments

16 citations


Proceedings ArticleDOI
08 Feb 2007
TL;DR: In this article, the authors have developed advanced CMOS pixel designs for scientific applications, such as Vanilla, a 520x520 array of 25μm pixels benefits from flushed reset circuitry for low noise and random pixel access for region of interest (ROI) readout.
Abstract: A UK consortium (MI3) has been founded to develop advanced CMOS pixel designs for scientific applications. Vanilla, a 520x520 array of 25μm pixels benefits from flushed reset circuitry for low noise and random pixel access for region of interest (ROI) readout. OPIC, a 64x72 test structure array of 30μm digital pixels has thresholding capabilities for sparse readout at 3,700fps. Characterization is performed with both optical illumination and x-ray exposure via a scintillator. Vanilla exhibits 34±3e- read noise, interactive quantum efficiency of 54% at 500nm and can read a 6x6 ROI at 24,395fps. OPIC has 46±3e- read noise and a wide dynamic range of 65dB due to high full well capacity. Based on these characterization studies, Vanilla could be utilized in applications where demands include high spectral response and high speed region of interest readout while OPIC could be used for high speed, high dynamic range imaging.

3 citations