A time-resolved, spectroscopic, diffuse optical tomography device was assembled for clinical applications like brain
functional imaging. The entire instrument lies in a unique setup that includes a light source, an ultrafast time-gated
intensified camera and all the electronic control units. The light source is composed of four near infrared laser diodes
driven by a nanosecond electrical pulse generator working in a sequential mode at a repetition rate of 100 MHz. The
light pulses are less than 80 ps FWHM. They are injected in a four-furcated optical fiber ended with a frontal light
distributor to obtain a uniform illumination spot directed towards the head of the patient. Photons back-scattered by the subject are detected by the intensified CCD camera. There are resolved according to their time of flight inside the head. The photocathode is powered by an ultrafast generator producing 50 V pulses, at 100 MHz and a width corresponding to a 200 ps FWHM gate. The intensifier has been specially designed for this application. The whole instrument is controlled by an FPGA based module. All the acquisition parameters are configurable via software through an USB plug and the image data are transferred to a PC via an Ethernet link. The compactness of the device makes it a perfect device for bedside clinical applications. The instrument will be described and characterized. Preliminary data recorded on test samples will be presented.
Several works have demonstrated the successfully integration of Single-photon avalanche photodiodes (SPADs)
operating in Geiger mode in a standard CMOS circuit for the last 10 years. These devices offer an exceptional temporal
resolution as well as a very good optical sensitivity. Nevertheless, it is difficult to predict the expected performances of
such a device. Indeed, for a similar structure of SPAD, some parameter values can differ by two orders of magnitude
from a technology to another. We proposed here a procedure to identify in just one or two runs the optimal structure of
SPAD available for a given technology. A circuit with an array of 64 SPAD has been realized in the Tower-Jazz 0.18 μm
CMOS image sensor process. It encompasses an array of 8 different structures of SPAD reproduced in 8 diameters in the
range from 5 μm up to 40 μm. According to the SPAD structures, efficient shallow trench insulator and/or P-Well guard
ring are used for preventing edge breakdown. Low dark count rate of about 100 Hz are expected thanks to the use of
buried n-well layer and a high resistivity substrate. Each photodiode is embedded in a pixel which includes a versatile
quenching circuitry and an analog output of its cathode voltage. The quenching system is configurable in four operation
modes; the SPAD is disabled, the quenching is completely passive, the reset of the photodiode is active and the
quenching is fully active. The architecture of the array makes possible the characterization of every single photodiode
individually. The parameters to be measured for a SPAD are the breakdown avalanche voltage, the dark count rate, the
dead time, the timing jitter, the photon detection probability and the after-pulsing rate.
The paper describes the realization of a complete optical imaging device to clinical applications like brain functional
imaging by time-resolved, spectroscopic diffuse optical tomography. The entire instrument is assembled in a unique
setup that includes a light source, an ultrafast time-gated intensified camera and all the electronic control units. The light
source is composed of four near infrared laser diodes driven by a nanosecond electrical pulse generator working in a
sequential mode at a repetition rate of 100 MHz. The resulting light pulses, at four wavelengths, are less than 80 ps
FWHM. They are injected in a four-furcated optical fiber ended with a frontal light distributor to obtain a uniform
illumination spot directed towards the head of the patient. Photons back-scattered by the subject are detected by the
intensified CCD camera; there are resolved according to their time of flight inside the head. The very core of the
intensified camera system is the image intensifier tube and its associated electrical pulse generator. The ultrafast
generator produces 50 V pulses, at a repetition rate of 100 MHz and a width corresponding to the 200 ps requested gate.
The photocathode and the Micro-Channel-Plate of the intensifier have been specially designed to enhance the
electromagnetic wave propagation and reduce the power loss and heat that are prejudicial to the quality of the image. The
whole instrumentation system is controlled by an FPGA based module. The timing of the light pulses and the
photocathode gating is precisely adjustable with a step of 9 ps. All the acquisition parameters are configurable via
software through an USB plug and the image data are transferred to a PC via an Ethernet link. The compactness of the
device makes it a perfect device for bedside clinical applications.
The conventional streak camera (CSC) is an optoelectronic instrument which captures the spatial distribution versus time
of a ultra high-speed luminous phenomena with a picosecond temporal resolution and a typical spatial resolution of
60 μm. This paper presents two Integrated Streak Camera (ISC) architectures called MISC (M for Matrix) and VISC (V
for Vector) which replicate the functionality of a streak camera on a single CMOS chip.
The MISC structure consists of a pixel array, where the column depth together with the sampling rate determine the
observation window. For proper operation, the image of the slit has to be spread uniformly over the rows of the imager.
The VISC architecture is based on a single column of photosensors, where each element is coupled to a front-end and a
multi-sampling and storage unit. The observation window is determined by the sampling rate and the depth of the
memory frame. The measurement of a 6 ns FWHM 532 nm light pulse laser is reported for both ISCs. For the two
architectures, the spatial resolution is linked to the size and the number of the photodetectors.
The classical streak cameras use a vacuum tube making thus fragile, cumbersome and expensive. The FAst MOS Imager
(FAMOSI) project consists in reproducing completely this streak camera functionality with a single CMOS chip. The
advantages of on-chip functionalities lead to a power reduction, a lower cost and miniaturization. In this paper, we show
the capabilities of a prototype fabricated in the AMS 0.35 μm CMOS process. The chip is composed of 64 columns per
64 rows of pixels. The pixels have a size of 20 μm per 20 μm and a fill factor of 47 %. The Chip FAMOSI implements an
electronic shutter and an analog accumulation capability inside the pixel. With this pixel architecture, the sensor can
work in single shot mode when the light pulse power is sufficient and in repetitive mode, i.e. it can measure a recurrent
light pulse and accumulates the successive photo charges into an internal node, for low light pulse detection. This
repetitive mode utilizes an analog accumulation in order to improve the sensitivity and the signal to noise ratio of the
system. Characterizations under static and uniform illumination in single shot mode have been done in order to evaluate
the performances of the detector. The main noises levels have been evaluated and the experiments show that a
conversion gain of 4.8 μV/e- is obtained with a dynamic range of 1.2V. Moreover, the charge transfer characterization in
single shot mode has been realized. It permits to know which potential must be apply to the charge spill transistor to
obtain the whole dynamic of the output with a maximal transfer gain, what is primordial to optimize the analog
accumulation. Finally, the dynamic operation of the sensors is exposed. Measurements show a sample time of 715 ps and
a time resolution better than 2 ns. A 6 ns light pulse has been measured in single shot and in accumulation mode.
High speed cameras use the interesting performances of CMOS imagers which offer advantages in on-chip functionalities, system power reduction, cost and miniaturization. The FAst MOS Imager (FAMOSI) project consists in reproducing the streak camera functionality with a CMOS imager. In this paper, we present a new imager called FAMOSI 2 which implements an electronic shutter and analog accumulation capabilities inside the pixel. With this kind of pixel and the new architecture for controlling the integration, FAMOSI 2 can work in repetitive mode for low light power and in single shot mode for higher light power. This repetitive mode utilizes an analog accumulation to improve the sensitivity of the system with a standard Nwell/Psub photodiode. The prototype has been fabricated in the AMS 0.35 μm CMOS process. The chip is composed of 64 columns per 64 rows of pixels. The pixels have a size of 20 μm per 20 μm and a fill factor of 47 %. Characterizations under static and uniform illumination in single shot mode have been done in order to evaluate the performances of the detector. The main noises levels have been evaluated and the experiments show that a conversion gain of 4.8 μV/e- is obtained with a dynamic range of 1.2 V. Moreover, the charge transfer characterization in single shot mode has been realized. It permits to know which potential must be apply to the charge spill transistor to obtain the whole dynamic of the output with a maximal transfer gain, what is primordial to optimize the analog accumulation.
We designed a camera based on a fast CMOS APS imager for high speed optical detection which produces images simi-larly as a streak camera. This imager produces the intensity information I as function of one spatial dimension and time (I=f(x,t)) from one frame with two spatial dimensions. The time sweeping is obtained by delaying successively the integration phase for each pixel of the same row. For the first FAMOSI (Fast MOs Imager) prototype the start of in-tegration is given by the camera itself. This signal is injected to a laser trigger. This laser emits a 10 nanoseconds light pulse onto the sensor. The temporal evolution of the light pulse is then resolved by the camera with a resolution of 800 ps. In single shot, the maximum dynamic of the camera is estimated to 64 dB and is limited by the readout noise. We decide to work in accumulation mode in order to increase the signal to noise ratio of the camera. But the high laser trigger (about 20 ns rms) does not allow accumulation of several optical events without a large spreading. The camera has been modified in order to be triggered by an external signal delivered by a trigger unit. In this new configuration the laser emit pulses at a repetition rate of 50 Hz. A photodiode detect a part of the laser pulse and generate the trigger signal for FAMOSI. The laser pulse is delayed with an optical fibre before being directed to the camera. The trigger jitter obtained is then less than 100 ps and allows accumulation without significant loss of the temporal resolution. With accumulation the readout noise is attenuated by a √N factor. Then with N = 1000 accumulations, the dynamics approach 93 dB. This allows the camera to work similarly as a synchroscan streak camera and then to observe weak signal.
In this paper, we describe a low cost pulse generator based on bipolar transistors designed to emit very short light pulses with a simple laser diode. Its main characteristics are laser pulses as short as 60 ps Full Width at Half Maximum (FWHM) and a repetition rate from DC up to 100 MHz. The circuit is based on three standard RF bipolar transistors which emit a short electrical pulse of 1.2 ns FWHM with 5 Volt amplitude peak-peak into 50 Ω. This unit can directly drive a classical laser diode to emit short laser pulses. A DC current generator is added in order to polarize the laser diode near its emission point and reduce the power delivered by the pulse generator and enhance its life time. This generator allows the adjustment of the emitted power of the laser diode. If the power emitted is too high, secondary pulses are observed. There is an optimal power where there is a single short pulse with duration down to 60 ps FWHM. The system is simply triggered by a TTL signal and its structure allows the generation of a single shoot, or a complex pattern generation and repetitive generation. The unique limitation is the delay between pulse emissions which must be greater than 10 ns. The typical optical power with a standard laser diode at 660 nm (45 mW continuous) is 1.7 mW with a repetition rate of 80 MHz and a pulse width down to 60 ps measured with a synchroscan streak camera. The peak power is about 350 mW. The pulse to pulse jitter is less than 10 ps rms. This system can be a competitive alternative to expensive commercial products.
Nowadays, imagers based on CMOS active pixel sensors (APS) have performances that are competitive with those based on charge-coupled devices (CCD). CMOS imagers offer advantages in on-chip functionalities, system power reduction, cost and miniaturisation. The FAst MOS Imager (FAMOSI) project consists in reproducing the streak camera functionality with a CMOS imager. In this paper, we present the second version of FAMOSI which makes up for the drawbacks of the first one. FAMOSI 2 has a new architecture of pixel which implements an electronic shutter and analogue accumulation capabilities. With this kind of pixel and the new architecture for controlling the integration, FAMOSI 2 can work in the low power repetitive synchroscan mode. The prototype has been fabricated in the AMS 0.35μm CMOS process. The chip is composed of 64 columns per 64 rows of pixels. The pixels have a size of 20μm per 20μm and a fill factor of 47%. The simulation shows that a conversion gain of 3.4μV/e- is obtained with a dynamic range of 1.2V, a time resolution of 400ps and a light pulse repetitive rate of 300kHz.
We describe a method to increase the speed of convergence for the simultaneous reconstruction of absorption and scattering images in Diffuse Optical Tomography (DOT). We used the diffusion approximation of the radiative transfer equation and the Finite Element Method (FEM) to solve the forward problem. The absorption and reduced scattering images are reconstructed by inverting the distribution of the moments of the time-dependent detected light flux. The inverse problem is solved with an optimization algorithm like ART or Conjugate Gradient. This ill-posed inverse problem can be simplified by using a priori knowledge of the studied objects.
In this paper, we consider that DOT is a functional imaging technique that can be complemented by an anatomical imaging technique like Magnetic Resonance Imaging (MRI). We used anatomic information obtained from MRI as prior knowledge to compute optical absorption and scattering images. In a first step, MRI segmented images were only
used to mesh our phantoms, with a finer resolution around boundaries. In a second step, we computed optical images with homogeneous properties from the segmented MRI image, in order to initialise our optimisation process. These two initialisations yield better reconstructed images. Reconstruction from simulated and experimental data will be presented.
Different temporal instabilities, which degrade the temporal resolution of s a synchroscan streak camera, have been studied. Each of the 3 main components: the laser, the trigger and the streak camera, have their intrinsic instability, thus a degradation of the final temporal resolution is occurred. An internal PLL in the streak camera has been developed in order to improve the temporal resolution. The synchroscan signal is used to lock the phase of the deflection voltage with the laser beam as close as possible. The phase detector has 0 to 360° area detection and a jitter lower than 300 fs FWHM integrated from 10 to 600 kHz, allowing sub picoseconds synchronization with the laser beam. The slow drifts, from 0 to 25 Hz, of the phase comparator are cancelled with a laser reference directly inserted in the camera input. By the way of an image processing, the phase command voltage is modified to lock the position of this laser reference. Results show that this stabilized camera can be used immediately after it is turned on (suppression of the warm-up time) and has very good temporal resolution, even with a long time exposure (2.4 ps FWHM with a time exposure of 2 hours has been realized). This allows more exploration in detection of very weak signals.
We describe a method to increase the convergence speed for simultaneous reconstruction of absorption and scattering images in Diffuse Optical Tomography (DOT). We used the diffusion approximation of the radiative tranfer equation and the Finite Element Method (FEM) to solve the forward problem. The absorption and reduced scattering images were reconstructed by inverting the distribution of the moments of the time-dependent detected light flux. The inverse problem is solved with an optimization algorithm such as ART or Conjugate Gradient. This ill-posed inverse problem can be simplified by using a priori knowledge of the objects studied.
In this paper, we consider that DOT is a functional imaging technique that can be complemented by an anatomical imaging technique such as Magnetic Resonance Imaging (MRI). The algorithm is built as follows: the first step consists in to dividing the observed zone into regions, thanks to another imaging technique such as MRI. In practice, this allows to adapt the mesh to the internal geometry. Then, supposing that each region is homogeneous absorption- and scattering-wise, a few parameters are reconstructed with an optimization technique. With a few iterative steps, well-averaged parameters can be obtained, which could be used to initialize the first stage of a global process. This process could reconstruct smaller inhomogeneities. We compare this method with direct global reconstruction, beginning with homogeneous parameters.
Two experimental setups for time-resolved diffuse optical tomography (DOT) are described. The first involves a titanium-sapphire laser and a streak camera in a synchroscan mode. A multiple-arm light guide allows to measure simultaneously the temporal profiles of photons re-emitted at different boundary sites of the objects studied. The second one uses picosecond laser diodes and a multiple-anode micro-channel plate photomultiplier tube followed by parallel time-correlated single photon counting channels. The two instruments are described and their main specifications compared: instrument response and time resolution, sweep time, temporal stability, sensitivity and its inter-channel variation, cross-talk between adjacent channels. We will also discuss detector-specific data analysis prior to image reconstruction. Both instruments have been tested with phantoms simulating tissues and the absorption and reduced scattering images obtained have been compared.
A near-infrared optical tomograph has been developed to obtain 3-dimensional images of scattering phantoms with optical properties similar to those of biological tissues. This experimental setup uses a femtosecond laser and a synchroscan streak camera. Thanks to three stepping motors, the phantom to be imaged is scanned in a parallel-beam mode. The time resolved detection of scattered photons is performed in the transmission mode, for different phantom positions obtained by two translations and one rotation stages. Regarding the data processing, we have developed a nonlinear image reconstruction algorithm based on the Newton-Raphson iterative method. It executes absorption and scattering mapping on the basis of characteristic data extracted from the recorded temporal point spread function of light transmitted through the studied object, such as the mean flight time of photon, the related variance or the integral intensity. First, 2D images using absolute or differential imaging schemes have been obtained for different scattering cylindrical phantoms possessing one or two more absorbing and/or more scattering inclusions. Reconstructed images have been compared to simulated and expected values. The results demonstrate that this system is a reliable and valuable platform for research on time-resolved optical tomography.
An experimental system for time-resolved optical tomography has been constructed from Ti:Sapphire laser and synchroscan streak camera. With a 3D step-motor driving translation- rotation stage, the system can scan the object to be imaged in a parallel-beam mode analogous to X-ray CT, by which the time-resolved projections at a variety of incident angles can be obtained automatically. By applying a nonlinear image reconstruction algorithm, the promising images have been produced from the time-resolved measurements for three kinds of tissue-like phantoms that represent the different combinations of the absorption and scattering properties of the embedded single inclusion. Our results have demonstrated this system as a reliable and ideal platform for the research of time-resolved tomographic imaging.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.