Automatic analysis of neuronal structure from wide-field-of-view 3D image stacks of retinal neurons is essential for
statistically characterizing neuronal abnormalities that may be causally related to neural malfunctions or may be early
indicators for a variety of neuropathies. In this paper, we study classification of neuron fields in large-scale 3D confocal
image stacks, a challenging neurobiological problem because of the low spatial resolution imagery and presence of
intertwined dendrites from different neurons. We present a fully automated, four-step processing approach for neuron
classification with respect to the morphological structure of their dendrites. In our approach, we first localize each
individual soma in the image by using morphological operators and active contours. By using each soma position as a
seed point, we automatically determine an appropriate threshold to segment dendrites of each neuron. We then use
skeletonization and network analysis to generate the morphological structures of segmented dendrites, and shape-based
features are extracted from network representations of each neuron to characterize the neuron. Based on qualitative
results and quantitative comparisons, we show that we are able to automatically compute relevant features that clearly
distinguish between normal and abnormal cases for postnatal day 6 (P6) horizontal neurons.
The use of radiation sensors as portal monitors is increasing due to heightened concerns over the smuggling of fissile
material. Transportable systems that can detect significant quantities of fissile material that might be present in vehicular
traffic are of particular interest, especially if they can be rapidly deployed to different locations. To serve this
application, we have constructed a rapid-deployment portal monitor that uses visible-light and gamma-ray imaging to
allow simultaneous monitoring of multiple lanes of traffic from the side of a roadway. The system operation uses
machine vision methods on the visible-light images to detect vehicles as they enter and exit the field of view and to
measure their position in each frame. The visible-light and gamma-ray cameras are synchronized which allows the
gamma-ray imager to harvest gamma-ray data specific to each vehicle, integrating its radiation signature for the entire
time that it is in the field of view. Thus our system creates vehicle-specific radiation signatures and avoids source
confusion problems that plague non-imaging approaches to the same problem. Our current prototype instrument was
designed for measurement of upto five lanes of freeway traffic with a pair of instruments, one on either side of the
roadway. Stereoscopic cameras are used with a third "alignment" camera for motion compensation and are mounted on
a 50' deployable mast. In this paper we discuss the design considerations for the machine-vision system, the algorithms
used for vehicle detection and position estimates, and the overall architecture of the system. We also discuss system
calibration for rapid deployment. We conclude with notes on preliminary performance and deployment.
Segmentation, tracking, and tracing of neurons in video imagery are important steps in many neuronal migration
studies and can be inaccurate and time-consuming when performed manually. In this paper, we present an
automated method for tracing the leading and trailing processes of migrating neurons in time-lapse image stacks
acquired with a confocal fluorescence microscope. In our approach, we first locate and track the soma of the
cell of interest by smoothing each frame and tracking the local maxima through the sequence. We then trace
the leading process in each frame by starting at the center of the soma and stepping repeatedly in the most
likely direction of the leading process. This direction is found at each step by examining second derivatives of
fluorescent intensity along curves of constant radius around the current point. Tracing terminates after a fixed
number of steps or when fluorescent intensity drops below a fixed threshold. We evolve the resulting trace to
form an improved trace that more closely follows the approximate centerline of the leading process. We apply a
similar algorithm to the trailing process of the cell by starting the trace in the opposite direction. We demonstrate
our algorithm on two time-lapse confocal video sequences of migrating cerebellar granule neurons (CGNs). We
show that the automated traces closely approximate ground truth traces to within 1 or 2 pixels on average.
Additionally, we compute line intensity profiles of fluorescence along the automated traces and quantitatively
demonstrate their similarity to manually generated profiles in terms of fluorescence peak locations.
The use of radiation sensors as portal monitors is increasing due to heightened concerns over the smuggling of fissile
material. Portable systems that can detect significant quantities of fissile material that might be present in vehicular
traffic are of particular interest. We have constructed a prototype, rapid-deployment portal gamma-ray imaging portal
monitor that uses machine vision and gamma-ray imaging to monitor multiple lanes of traffic. Vehicles are detected
and tracked by using point detection and optical flow methods as implemented in the OpenCV software library. Points
are clustered together but imperfections in the detected points and tracks cause errors in the accuracy of the vehicle
position estimates. The resulting errors cause a "blurring" effect in the gamma image of the vehicle. To minimize these
errors, we have compared a variety of motion estimation techniques including an estimate using the median of the
clustered points, a "best-track" filtering algorithm, and a constant velocity motion estimation model. The accuracy of
these methods are contrasted and compared to a manually verified ground-truth measurement by quantifying the rootmean-
square differences in the times the vehicles cross the gamma-ray image pixel boundaries compared with a groundtruth
manual measurement.
Many composite correlation filter designs have been proposed for solving a wide variety of target detection and pattern recognition problems. Due to the large number of available designs, however, it is often unclear how to select the best design for a particular application. We present a theoretical survey and an empirical comparison of several popular composite correlation filter designs. Using a database of rotational target imagery, we show that some such filter designs appear to be better choices than others under computational and performance constraints. We compare filter performance in terms of noise tolerance, computational load, generalization ability, and distortion in order to provide a multifaceted examination of the characteristics of various filter designs.
Most integrated target detection and tracking systems employ state-space models to keep track of an explicit
number of individual targets. Recently, a non-state-space framework was developed for enhancing target detection
in video by applying probabilistic motion models to the soft information in correlation outputs before
thresholding. This framework has been referred to as multi-frame correlation filtering (MFCF), and because it
avoids the use of state-space models and the formation of explicit tracks, the framework is well-suited for handling
scenes with unknown numbers of targets at unknown positions. In this paper, we propose to use quadratic
correlation filters (QCFs) in the MFCF framework for robust target detection. We test our detection algorithm
on real and synthesized single-target and multi-target video sequences. Simulation results show that MFCF can
significantly reduce (to zero in the best case) the false alarm rates of QCFs at detection rates above 95% in the
presence of large amounts of uncorrelated noise. We also show that MFCF is more adept at rejecting those false
peaks due to uncorrelated noise rather than those due to clutter and compression noise; consequently, we show
that filters used in the framework should be made to favor clutter rejection over noise tolerance.
Distortion-tolerant correlation filter methods have been applied to many video-based automatic target recognition (ATR) applications, but in a single-frame architecture. In this paper we introduce an efficient framework for combining information from multiple correlation outputs in a probabilistic way. Our framework is capable of handling scenes with an unknown number of targets at unknown positions. The main algorithm in our framework uses a probabilistic mapping of the correlation outputs and takes advantage of a position-independent target motion model in order to efficiently compute
posterior target location probabilities. An important feature of the framework is the ability to incorporate any existing correlation filter design, thus facilitating the construction of a distortion-tolerant multi-frame ATR. In our simulations, we incorporate the minimum average correlation energy Mellin radial harmonic (MACE-MRH) correlation filter design, which allows the user to specify the desired scale response of the filter. We test our algorithm on real and synthesized infrared (IR) video sequences that exhibit various degrees of target scale distortion. Our simulation results show that the multi-frame algorithm significantly improves the recognition performance of a MACE-MRH filter while requiring only a marginal increase in computation. We also show that, for an equivalent amount of added computation, using larger filter banks instead of multi-frame information is unable to provide a comparable performance increase.
Correlation filters are attractive for automatic target recognition (ATR) applications due to such attributes as shift invariance, distortion tolerance and graceful degradation. Composite correlation filters are designed to handle target distortions by training on a set of images that represent the expected distortions during testing. However, if the distortion can be described algebraically, as in the case of in-plane rotation and scale, then only one training image is necessary. A recently introduced scale-tolerant correlation filter design, called the Minimum Average Correlation Energy Mellin Radial Harmonic (MACE-MRH) filter, exploits this algebraic property and allows the user to specify the scale response of the filter. These filters also minimize the average correlation energy in order to help control the sidelobes in the correlation output and produce sharper, more detectable peaks. In this paper we show that applying non-linearities in the frequency domain (leading to fractional power scale-tolerant correlation filters) can significantly improve the resulting peak sharpness, yielding larger peak-to-correlation energy values for true-class targets at various scales in a scene image. We investigate the effects of fractional power transformations on MACE-MRH filter performance by using a testbed of fifty video sequences consisting of long-wave infrared (LWIR) imagery, in which the observer moves along a flight path toward one or more ground targets of interest. Targets in the test sequences suffer large amounts of scale distortion due to the approach trajectory of the camera. MACE-MRH filter banks are trained on single targets and applied to each sequence on a frame-by-frame basis to perform target detection and recognition. Recognition results from both fractional power MACE-MRH filters and regular MACE-MRH filters are provided, showing the improvement in scale-tolerant recognition from applying fractional power non-linearities to these filters.
Robust real-time recognition of multiple targets with varying pose requires heavy computational loads, which are often too demanding to be performed online at the sensor location. Thus an important problem is the performance of ATR algorithms on highly-compressed video sequences transmitted to a remote facility. We investigate the effects of H.264 video compression on correlation-based recognition algorithms. Our primary test bed is a collection of fifty video sequences consisting of long-wave infrared (LWIR) and mid-wave infrared (MWIR) imagery of ground targets. The targets are viewed from an aerial vehicle approaching the target, which introduces large amounts of scale distortion across a single sequence. Each sequence is stored at seven different levels of compression, including the uncompressed version. We employ two different types of correlation filters to perform frame-by-frame target recognition: optimal tradeoff synthetic discriminant function (OTSDF) filters and a new scale-tolerant filter called fractional power Mellin radial harmonic (FPMRH) filters. In addition, we apply the Fisher metric to compressed target images to evaluate target class separability and to estimate recognition performance as a function of video compression rate. Targets are centered and cropped according to ground truth data prior to separability analysis. We compare our separability estimates with the actual recognition rates achieved by the best correlation filter for each sequence. Numerical results are provided for several target recognition examples.
High-resolution single photon emission computed tomography (SPECT) and X-ray computed tomography (CT) imaging have proven to be useful techniques for non-invasively monitoring mutations and disease progression in small animals. A need to perform in vivo studies of non-anesthetized animals has led to the development of a small-animal imaging system that integrates SPECT imaging equipment with a pose-tracking system. The pose of the animal is monitored and recorded during the SPECT scan using either laser-generated surfaces or infrared-reflective markers affixed to the animal. The reflective marker method measures motion by stereoscopically imaging an arrangement of illuminated markers. The laser-based method is proposed as a possible alternative to the reflector method with the advantage that it is a non-contact system. A three-step technique is described for calibrating the surface acquisition system so that quantitative surface measurements can be obtained. The acquired surfaces can then be registered to a reference surface using the iterative closest point (ICP) algorithm to determine the relative pose of the live animal and correct for any movement during the scan. High accuracy measurement results have been obtained from both methods.
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.