This PDF file contains the front matter associated with SPIE Proceedings Volume 8134, including the Title Page, Copyright information, Table of Contents, and Conference Committee listing.

Spike processing is one kind of hybrid analog-digital signal processing, which has the efficiency of analog processing and the robustness to noise of digital processing. When instantiated with optics, a hybrid analog-digital processing primitive has the potential to be scalable, computationally powerful, and have high operation bandwidth. These devices open up a range of processing applications for which electronic processing is too slow. Our approach is based on a hybrid analog/digital computational primitive that elegantly implements the functionality of an integrate-and-fire neuron using a Ge-doped non-linear optical fiber and off-the-shelf semiconductor devices. In this paper, we introduce our photonic neuron architecture and demonstrate the feasibility of implementing simple photonic neuromorphic circuits, including the auditory localization algorithm of the barn owl, which is useful for LIDAR localization, and the crayfish tail-flip escape response.

An autonomous wireless sensor network that consists of different types of sensor modalities is a topic of intense research due to its versatility and portability of applications. Typical autonomous sensor networks commonly include passive sensor nodes such as infrared, acoustic, seismic, and magnetic. However, fusion of another active sensor such as Doppler radar in the integrated sensor network may offer powerful capabilities for many different sensing and classification tasks. In this work, we demonstrate the design and implementation of an autonomous wireless sensor network integrating a Doppler sensor into wireless sensor node with commercial off the shelf components. Then we demonstrate two applications of the newly integrated radar mote in a wireless sensor network environment where other sensor motes are supporting the integrated radar mote for autonomous triggering and data collection. At first we use the integrated system to detect the range and velocity of a toy train effectively to demonstrate its capability as a surveillance tool. Then we classify different types of non-conducting target materials based on their reflected signal response to newly built radar mote. Different types of materials can usually affect the amount of energy reflected back to the source of an electromagnetic wave. For investigating this observation we simulate models for the reflectivity of different homogeneous non-conducting materials using a mathematical model and later classify the types of target materials using real experimental data collected through our autonomous radar-mote sensor network.

An iterative algorithm to design optimal trade-off correlation filters for pattern recognition is presented. The algorithm is based on a heuristic optimization of several conflicting metrics simultaneously. By the use of the heuristic algorithm the impulse response of a conventional composite filter is iteratively synthesized until an optimal trade-off of the considered quality metrics is obtained. Computer simulation results obtained with the proposed filters are provided and analyzed in terms of recognition quality measures in cluttered, and geometrically distorted input test scenes.

We have developed a novel face recognition technique utilizing optical joint transform correlation (JTC) technique which provides with a number of salient features as compared to similar other digital techniques, including fast operation, simple architecture and capability of updating the reference image in real time. The proposed technique incorporates a synthetic discriminant function (SDF) of the target face estimated from a set of different training faces to make the face recognition performance invariant to noise and distortion. The technique then involves four different phase-shifted versions of the same SDF reference face, which are individually joint transform correlated with the given input scene with unknown faces and other objects. Appropriate combination of correlation signals yields a single cross-correlation peak corresponding to each potential face image. The technique also involves a fringe-adjusted filter to generate a delta-like correlation peak with high discrimination between the target face and the non-target face and background objects. Performance of the proposed face recognition technique is investigated through computer simulation where it is observed to be efficient and successful in different complex environments.

In this paper we examine the acceleration of two spiking neural network models on three clusters of multicore processors representing three categories of processors: x86, STI Cell, and NVIDIA GPGPUs. The x86 cluster utilized consists of 352 dualcore AMD Opterons, the Cell cluster consists of 320 Sony Playstation 3s, while the GPGPU cluster contains 32 NVIDIA Tesla S1070 systems. The results indicate that the GPGPU platform can dominate in performance compared to the Cell and x86 platforms examined. From a cost perspective, the GPGPU is more expensive in terms of neuron/s throughput. If the cost of GPGPUs go down in the future, this platform will become very cost effective for these models.

Margin Setting is very successful way to do pattern recognition. Because it differs fundamentally from prior work it has proved difficult for me to understand even though I invented it years ago and even though it has proved easy to apply and extremely effective. It took me years to understand what I was doing in any fundamental way. Now that I have begun to understand it, the mystery of its success is removed. Its full effects are not well understood, so there is still important basic research to be done. And, the advantages of Margin Setting over prior methods have proved substantial. There seem to me no remaining barriers to its widespread use.

Normal hearing persons are able to localize the direction of sounds better using both ears than when listening with only one ear. Localization ability is dependent on auditory system perception of interaural differences in time, intensity, and phase. Interaural timing differences (ITDs) provide information for locating direction of low and mid frequency sounds, while interaural level differences (ILDs), which occur because of the horizontal plane shadowing effect of the head, provide information for locating direction of higher frequency sounds. The head related transfer function (HRTF) contains characteristic information important for acoustic localization. Models based on HRTFs take into account head shadow, torso, and pinna effects, and their impact on interaural frequency, level, and timing differences. Cochlear implants (CIs) have proven a successful treatment for persons with bilateral severe to profound hearing loss. A problem is that only some ITD and ILD cues are maintained with CI sound processing, and the microphone position alters the acoustic cues. The relative impact of differences in physical cues received by the auditory system with bilateral CIs versus differences in the ability of the damaged auditory nervous system to process bilateral inputs is not yet clear. The model presented in this paper was constructed as a step toward answering this question, and is intended to serve as a tool for future development of more optimal signal processing algorithms that may provide better localization ability for persons with bilateral CIs.

Computational reacquisition for real 3D object in integral imaging is proposed. The pseudoscopic problem and matching of lens array in pickup and display are fundamental problems for real-time broadcasting based on integral imaging. We propose the reconstruction method of real 3D object and computational reacquisition method without matching of pickup and display lens array. In reconstruction, the real 3D object is reconstructed with volumetric information using depth extraction and triangular mesh, which can be rotated and translated in orthoscopic geometry. In reacquisition, the virtual lens array can generate the elemental image without matching of pickup lens array using OpenGL.

The parallel spectrum analyzer of optical signals is presented. This device is one of spectral devices which receive spectroscopic information in an optical range. The method of receiving spectroscopic information based on the resonance phenomenon is described. The novelty of the method lies in implementing the spectral decomposition by the principle of the narrow-band optical filtration in n parallel channels of the spectrum analyzer. Furthermore, this device can make parallel analysis of spectrum of optical radiations which are in places of difficult access, for instance, in unfavorable conditions of high humidity, high temperature and toxic contamination. The result is achieved due to specific structure of the analyzer which contains n channels of spectrum analyzer and the group of optical fibers which transmits analyzed optical radiation on the given distance.

Intrinsic optical signal imaging (IOS) promises a noninvasive method for high resolution examination of retinal function. Using freshly isolated animal retinas, we have conducted a series of experiments to test fast IOSs which have time courses comparable to electrophysiological kinetics. In this article, we demonstrate the feasibility of in vivo imaging of fast IOSs correlated with retinal activation in anesthetized frog (Rana Pipiens). A rapid (68,000 lines/s) line-scan confocal ophthalmoscope was constructed to achieve high-speed (200 frames/s) near infared (NIR) recording of fast IOSs. By rejecting out-of-focus background light, the line-scan confocal imager provided enough resolution to differentiate individual photoreceptors in vivo. With visible light stimulation, NIR confocal images disclosed transient IOSs with time courses comparable to retinal ERG kinetics. High-resolution IOS images revealed both positive (increasing) and negative (decreasing) light responses, with sub-cellular complexity, in the activated retina.

We propose a technique that employs single fluorescent molecules for visualizing the distribution of strain induced in microstructures. We sprayed single-molecule tracers on microstructures by ultrasonic atomization and traced the position and orientation of the tracers by a single-molecule detection technique with a three-dimensional (3D) orientation microscope, which consists of a conventional fluorescent microscope and a polarization-mode converter. By using 3D spline interpolation, we visualized the surface geometry of a microelectromechanical (MEMS) device. We tracked the 3D position and orientation of tracers attached to a supporting beam of the MEMS mirror. The surface declination angles calculated from the orientation of the tracers were in agreement with the tilt angle obtained from the 3D position of the tracers.

We present the application of a simple physical model to accurately predict the broadband spectral transmittance and colorimetric properties of a twisted-nematic liquid crystal display (TNLCD). We spectroscopically calibrate the retardance parameters to evaluate the spectrum of the light transmitted by a TNLCD sandwiched between two linear polarizers. When the TNLCD is illuminated with a broadband light source, the full spectrum can be predicted as a function of the addressed grey level for any arbitrary orientation of the polarizers. Experimental results confirming the validity of the proposed systems are presented, both on the measured spectral responses as well as on the trajectories at different chromatic diagrams. The presented results can be useful for optoelectronic systems requiring an accurate control of the spectral characteristics of the light.

Quantitative phase microscopy with a closed-loop feedback control is described. The system is based on a Mach-Zehnder interferometer using a phase-locking technique and measures the change in optical path length while the sample is scanned across the optical axis. The important feature of this system is no need for any unwrapping process, which allows us to measure the sample with highly precise around 2π radian. The spatial resolution is estimated to be less than 1μm, which is limited by the Rayleigh criterion and the sample thickness. The path length resolution is estimated to be less than 1nm, which corresponds to the change in refractive index less than 10^-3 for the sample of 10μm thick. Precise measurements of quantitative phase images by a volume phase holographic grating or a multi-level zone plate array validate the principle and show the accuracy of the methods. The samples depending on wavelength with the pigment such as red blood cell and diatom cells are measured by different wavelength. We quantify its accuracy by measuring calibrated test samples. The technique is applicable to examination of limitation of passage culture, biological cell imaging for diagnosis.

It is practical and efficient to simplify targets to point scatterers in radar simulations. With low-resolution radars, the radar cross section (RCS) is a sufficient feature to characterize the scattering properties of a target. However, the RCS totals the target scattering properties to a scalar value for each aspect angle. Thus, a more detailed representation of the target is required with high-resolution radar techniques, such as Inverse Synthetic-Aperture Radar (ISAR). In straightforward simulation scenarios, high-resolution targets have been modeled placing identical point scatterers in the shape of the target, or with a few dominant point scatterers. As extremely simple arrangements, these do not take the self-shadowing into account and are not realistic enough for high demands. Our radar response simulation studies required a target characterization akin to RCS, which would also function in highresolution cases and take the self-shadowing and multiple reflections into account. Thus, we propose an approach to converting a 3-dimensional (3D) surface into a set of scatterers with locations, orientations, and directional scattering properties. The method is intended for far field operation, but could be adjusted for use in the near field. It is based on ray tracing which provides the self-shadowing and reflections naturally. In this paper, we present ISAR simulation results employing the proposed method. The constructed scatterer set is scalable for different wavelengths enabling the fast production of realistic simulations including authentic RCS scattering center formation. This paper contributes to enhancing the reality of the simulations, yet keeping them manageable and computationally reasonable.

Two ways to improve spatial coherence in high power beams are considered and compared, spatial filter and adaptive optics. High power in lasers is usually produced by using a succession of optical amplifiers. As intensity increases in optical amplifiers or lasers, spatial coherence decreases because of uneven heating and other nonlinearities. Often a spatial filter in the form of a pinhole is used after each stage that will block the parts of the beam that are spatially incoherent and would scatter sideways out of the beam. The loss of light blocked by the pinhole lowers amplification so that more amplifiers are required. An alternative approach to improving spatial coherence is adaptive optics. In this case the wavefront is measured using a wavefront sensor such as a Hartman wavefront sensor or with a photonic crystal. The measurements are used to compute settings for a deformable mirror. This involves a complicated feedback loop. The deformable mirror reduces the undesirable deviations in the wavefront to improve spatial coherence.

The 192 laser beams in the National Ignition Facility (NIF) are automatically aligned to the target-chamber center using images obtained through charged-coupled-device (CCD) cameras. Several of these cameras are in and around the target chamber during an experiment. Current experiments for the National Ignition Campaign are attempting to achieve nuclear fusion. Neutron yields from these high-energy fusion shots expose the alignment cameras to neutron radiation. The present work explores modeling and predicting laser alignment performance degradation due to neutron radiation effects, and introduces techniques to mitigate performance degradation. Camera performance models have been created based on the predicted camera noise from the cumulative neutron fluence at the camera location. We have found that the effect of the neutron-generated noise for all shots to date have been well within the alignment tolerance of half a pixel, and image processing techniques can be utilized to reduce the effect even further on the beam alignment to target.

To test the parameters of image intensifier screen is the precondition for researching and developing the third generation image intensifier. The picture of brightness uniformity of tested fluorescence screen shows bright in middle and dark at edge. It is not so direct to evaluate the performance of fluorescence screen. We analyze the energy and density distribution of the electrons, After correction, the image in computer is very uniform. So the uniformity of fluorescence screen brightness can be judged directly. It also shows the correction method is reasonable and close to ideal image. When the uniformity of image intensifier fluorescence screen brightness is corrected, the testing instrument is developed. In a vacuum environment of better than 1×10^-4Pa, area source electron gun emits electrons. Going through the electric field to be accelerated, the high speed electrons bombard the screen and the screen luminize. By using testing equipment such as imaging luminance meter, fast storage photometer, optical power meter, current meter and photosensitive detectors, the screen brightness, the uniformity, light-emitting efficiency and afterglow can be tested respectively. System performance are explained. Testing method is established; Test results are given.

In this paper, we propose a new technique for rotation invariant recognition and tracking of the face of a target person in a given scene. We propose an optimized method for face tracking based on the Fringe-adjusted JTC architecture. To validate our approach, we used the PHPID data base containing1 faces with various in-plane rotations. To enhance the robustness of the proposed method, we used a three-step optimization technique by: (1) utilizing the fringe-adjusted filter (H_FAF) in the Fourier plane, (2) adding nonlinearity in the Fourier plane after applying the H_FAF filter, and (3) using a new decision criterion in the correlation plane by considering the correlation peak energy and five largest peaks outside the highest correlation peak. Several tests were made to reduce the number of reference images needed for fast tracking while ensuring robust discrimination and efficient of the desired target.

To characterize the degree of consistency of parameters of the optically uniaxial birefringent protein nets of blood plasma a new parameter - complex degree of mutual anisotropy is suggested. The technique of polarization measuring the coordinate distributions of the complex degree of mutual anisotropy of blood plasma is developed. It is shown that statistic approach to the analysis of complex degree of mutual anisotropy distributions of blood plasma is effective in the diagnosis and differentiation of acute inflammation - acute and gangrenous appendicitis.

The paper is aimed at researching the effectiveness of differentiation of optical anisotropic property of biological crystal networks by means of the statistic and correlation analysis of new analytical parameter - the degree of mutual correlation of biological tissues Mueller matrices of different morphological structure and physiological state.

The depth resolution of the Optical Coherence Tomography (OCT) system is dependent on the coherence length and the lateral resolution depends on the beam waist on the sample. In OCT, optical axial resolution can thus be increased by selecting a broad band source. High NA lenses provide smaller spot size than low NA lens at the expense of short depth of focus that often does not stretch along. This is a trade-off between lateral resolution and depth of focus. In this paper, a new kind of polarization pupil mask, which consists of a three-portion half-wave plate between two parallel quarter-wave plates, is introduced into the Optical Coherence Tomography (OCT) systems, has the ability to realize the lateral superresolution and extend the focal depth in a small range when used as an apodizer in the optical imaging system. The simulation results show that high focal depth which is two times of the original optical system and the lateral superresolution can be realized at the same time.

In this paper, we propose a new method that can acquire the complex hologram of a 3D object for optical scanning holography (OSH) by using polarized optical elements and a 90 degrees 2×4 optical hybrid. The 3D object is scanned by a wave front combined by a plane and spherical wave with orthogonal polarization. The transmitted/reflected signal light is divided into two beams by a polarized beamsplitter (PBS). Then after the polarization control, the two separated beams are heterodyne detected by a 90 degrees 2×4 optical hybrid with two balanced detectors. The 3D object can be reconstructed from the complex hologram acquired from the balanced detectors. Compared with the conventional OSH, the acousto-optic modulator and complex electronic system are not needed in this method. Another advantage of this method lies in its high efficiency since there is almost no energy loss in the generation of the scanning beam by using the polarized optical elements. This method should be used for 3D holographic microscopy and 3D holographic television.

Fresnel telescopy full-aperture synthesized imaging ladar is a new high resolution active laser imaging technique. In the operational mode with moving target by beam scanning, spatial distributions of the complex data of the returned signals are not regular or uniform. In order to use fast Fourier transform (FFT) algorithm in signal processing, algorithm for Fresnel telescopy imaging system should include resampling interpolation step to change sampling data into uniform and quadratic spatial distribution. We choose suitable resampling interpolation algorithm for Fresnel telescopy, and get computer simulation of area target. The work is found to have substantial practical value and offers significant practical benefit for high resolution imaging in Fresnel telescopy imaging ladar.

Based on the computer vision theory, this article researched the algorithm for the location of laser beam cutting. This article combines Canny operator and thresheolding image. It overcomes the inaccuracy of edge detection and clutter jamming problem which are caused by the poor quality of acquired images. Collecting the key points of target edge, making B-spline curves fitting, it solved the problem that target edge is jagged. And it uses interpolation algorithm to locate the point for laser beam cutting. At last, we developed corresponding and prefessional software system which is based on Visual Stdio2003 and C#.

New kind of spectral processing of optical signals - the analysis of sample spectra is considered. These spectra are variable in time and introduced into the theory of spectral optical measurements for the first time. Spectral devices calculating sample spectra in optical range are considered as linear systems. Spectrum spread function of spectral device that performs the analysis of a sample of the spectrum is introduced as a reaction to the function. It is a linear integral operator kernel which establishes spectral device input-output ratio. Complex sample spectrum is obtained by using a dispersing system which is considered as optical four-pole network, a comprehensive description of which is the transfer function.

We propose and demonstrate a novel visual encryption device composed of higher-order birefringent elements. When an optical material with higher-order birefringence is placed between a pair of polarizers and illuminated by white light, it appears only white. In contrast, when it is illuminated by monochromatic light, the transmitted intensity varies depending cosinusoidally on the wavelength. An array of such materials can express information (e.g., letters and/or images) by controlling the birefringence of each pixel. If birefringence phase retardation can be adjusted for a specific wavelength, the information will be clearly displayed when it is illuminated at this wavelength. We denote this wavelength a key wavelength. The encryption device was fabricated by controlling the amount of higher-order birefringence to achieve high contrast only by using polarized illumination at the key wavelength. Thus, the information stored in the encryption device can be decoded only by illuminating it at the key wavelength. To demonstrate the validity of this encryption principle, we constructed a 3 × 3 pixel device in which commercial retarder films were laminated. The device was illuminated by a monochromatic light. When a readout experiment was performed using the monochromatic light at the key wavelength, the stored letter was clearly visible. On the other hand, when pixel brightness was randomly distributed with illumination at the wavelength other than the key wavelength, the letter could not be recognized. Furthermore, the stored information can be easily distributed to multiple physical keys that display arbitrary images. In this case, the birefringence phase retardation is obtained by summing the values of retardation of each pixel of the physical keys. In the experimental device, the observed image was decoded by superimposing the two images using different physical keys.