This PDF file contains the front matter associated with SPIE Proceedings Volume 6979, including the Title Page, Copyright information, Table of Contents, Introduction, and the Conference Committee listing.

Subband/Wavelet filter analysis-synthesis filters are a major component in many compression algorithms. Such compression algorithms have been applied to images, voice, and video. These algorithms have achieved high performance. Typically, the configuration for such compression algorithms involves a bank of analysis filters whose coefficients have been designed in advance to enable high quality reconstruction. The analysis system is then followed by subband quantization and decoding on the synthesis side. Decoding is performed using a corresponding set of synthesis filters and the subbands are merged together. For many years, there has been interest in improving the analysis-synthesis filters in order to achieve better coding quality. Adaptive filter banks have been explored by a number of authors where by the analysis filters and synthesis filters coefficients are changed dynamically in response to the input. A degree of performance improvement has been reported but this approach does require that the analysis system dynamically maintain synchronization with the synthesis system in order to perform reconstruction. In this paper, we explore a variant of the adaptive filter bank idea. We will refer to this approach as fixed analysis adaptive synthesis filter banks. Unlike the adaptive filter banks proposed previously, there is no analysis synthesis synchronization issue involved. This implies less coder complexity and more coder flexibility. Such an approach can be compatible with existing subband wavelet encoders. The design methodology and a performance analysis are presented.

The paper proposes an efficient iris recognition algorithm, obtained through the fusion of Haar Wavelet and Circular Mellin operator. The recognition system preprocesses the captured iris image to remove the effect of holes or spot of light lying on the pupillary region which creates problem in pupil localization. The processed image is localized by detecting inner and outer boundaries from the pupil center using maximum value of the spectrum image. Then the eyelids are detected by fitting a 3^rd degree polynomial on the suitable edge segments and removing the region occluded by eyelids from the normalized iris image. The features for the iris pattern are extracted using Haar Wavelet and Circular Mellin operator. The Haar Wavelet decomposition reduces the size of feature vector while Circular Mellin operator is used for rotation and scale invariant feature extraction. The features are compared using Hamming Distance method and the fusion is done at decision level using Conjunction rule. The recognizer is found to be more robust with accuracy level more than 95%.

The fingerprint recognition is the most developed field in the biometrics recognition and it can apply to the various applications. So the various studies are in progress for an enhancement of fingerprint recognition performance. In this paper, we will study the fingerprint recognition sensor. The fingerprint becomes input using fingerprint sensor to the fingerprint recognition system and the performance of the sensor has an effect on the fingerprint quality. Therefore, the improvement of system performance is possible by improving the performance of sensor. In this paper, we will study sensors which can be protected or unprotected from artificial fingerprint attack. We make various artificial fingerprints, and test sensors to overcome artificial fingerprint attack. We will analyze results of scanning test according to the various sensors and propose the method using the histogram of the normal image in order to measure the performance of these sensors. The measured characteristics are the resolution, shift, gradient, contrast, and rotation. The purpose of this paper is to propose the ways of the performance measurement which can be a criterion to evaluate the sensor performance.

To address the challenges on non-cooperative long-distance human authentication, identification, and verification; we propose an innovative scheme for developing a robust and automatic long-range biometric recognition system by combining face recognition and iris recognition of non-cooperative individuals in 24/7 operations. The system consists of three cameras. One is a wide field of view (WFOV) CCD video camera with InfraRed (IR) filter and powerful IR illuminators for human scan in a wide area and from a long distance. The other two cameras are high resolution video cameras with narrow field of view (NFOV) and IR filter & illuminators, mounted on a pan-tilt-unit (PTU) to capture the frontal view of human face and iris respectively. The WFOV detects the person and the NFOV cameras extract details for person identification. Once the frontal view shots are captured by the NFOV cameras, the face/iris models will be extracted by applying the state-of-the-art face/iris recognizers. In addition, a multimodality fusion approach integrates the face and iris recognition results to improve the overall recognition performance.

In this paper we describe design options when implementing sampling rate converters with FPGAs. We first review typical designs using IIR and FFT-based systems and then show implementations of fractional sampling rate changer ranging from Lagrange, B-spline to recently introduced C-MOMS and O-MOMS designs. Speed, area and error performance results for the circuits designed in VHDL are provided.

This is the third paper in a series that introduces a MatLab/Simulink-based design flow for FPGAs at an undergraduate curriculum level. In the first paper presented at SPIE 2006 we analyzed the design tools, while in the second paper presented at SPIE 2007 we reported on the appropriate topics for the lectures and labs. In this third paper we first give an overview based on the 12-year EDA cycle on why FPGAs have now reached a level where SOPC design is possible and why MatLab/Simulink is favored by both leaders in the FPGA field: Altera (DSP builder) and Xilinx (System Generator). We then describe the Xilinx Blackboard educational material development (EMD) that has been used in Spring 2007 and Spring 2008 to teach a Xilinx System Generator based course and laboratory.

In this paper we provide a simple and fast hardware implementation for a Support Vector Machine (SVM). By using the CORDIC algorithm and implementing a 2-based exponential kernel that allows us to simplify operations, we overcome the problems caused by too many internal multiplications found in the classification process, both while applying the Kernel formula and later on multiplying by the weights. We show a simple example of classification with the algorithm and analyze the classification speed and accuracy.

In this paper we address the problem of Support Vector Machine (SVM) learning. We describe an analogue implementation for a Sequential Minimal Optimization (SMO) algorithm to simplify the hardware requisites of the learning phase. The advantages from a full set training circuit are shown and a test is carried out on a simple case to prove its effectiveness.

We establish robust stability results for the bacterial heat stress response by using a reduced - order model with different time-scales under parameter perturbations and determine conditions that ensure the existence of asymptotically stable equilibria of the perturbed system. It is assumed that the system uncertainties are limited by the upper bounds of their norms. We derive a Lyapunov function for the coupled system and a maximal upper bound for the fast time scale associated with the unfolded proteins state.

Based on his early graduated studies in psychophysics, the author has, in recent years, applied psychophysics for studying organic and motor senses (the two sensory systems deeply embedded inside of human body), and tried to understand the scientific foundation of the oriental health promoting practices. The preliminary results are promising and are discussed in detail in this paper. Psychophysics studies of organic and motor senses may be the tool to provide the connection between Western and Eastern medicines to form a balanced holistic medicine approach, and may help us to understand the scientific foundation of mysterious oriental health Promoting practices that serve as alternative medicines for promoting human wellness against illness.

The complexity of gene regulatory networks described by coupled nonlinear differential equations is often an obstacle for analysis purposes. Therefore, the development of effective model reduction techniques is of paramount importance in the field of systems biology. In this paper, we apply the theory of nonlinear balanced truncation for model reduction for gene regulatory networks based only on standard matrix computations. The method is based on finding a controllability and observability function of the nonlinear system and thus obtain a balanced representation that produces singular value functions which are functions of the state. As a result, we obtain a ranked contribution of the states from an input - output perspective.

An application of dependent component analysis techniques is reported for the detection and characterization of small indeterminate breast lesions in dynamic contrast-enhanced MRI. These techniques enable the extraction of spatial and temporal features of dynamic MRI data stemming from patients with confirmed lesion diagnosis. By revealing regional properties of contrast-agent uptake characterized by subtle differences of signal amplitude and dynamics, this method provides both a set of prototypical time-series and a corresponding set of cluster assignment maps which further provides a segmentation with regard to identification and regional subclassification of pathological breast tissue lesions. We present two different segmentation methods for the evaluation of signal intensity time courses for the differential diagnosis of enhancing lesions inStarting from the conventional methodology, we proceed by introducing the separate concepts of threshold segmentation and dependent component analysis and in the last step by combining those two concepts. The results suggest that the dependent component approach has the potential to increase the diagnostic accuracy of MRI mammography by improving the sensitivity without reduction of specificity.

Nanotechnology is expected to provide the fundamental basis of the next two generations of products and processes. Impacts for applications are already being felt in many fields, and there is interest especially in the aerospace industry, where performance is a major driver of decisions for applications. Four areas are receiving special emphasis in a program aimed at the Air Force's strategic focus on materials. The emphasis includes adaptive coatings and surface engineering, nanoenergetics, electromagnetic sensors, and power generation and storage. Seven universities in Texas have initiated the CONTACT program of focused research including nine projects in the first year, with plans for expansion in subsequent years. This paper discusses the focus, progress, and plans for the second year and opportunities for industry input to the scope and content of the research. A new model for the creation and guidance of research programs for industry is presented. The new approach includes interaction with the aerospace industry and the Air Force that provides a focus for the research. Results to date for the new method and for the research are presented. A discussion of nanoengineering technology transition into the aerospace industry highlights the mechanisms for enhancing the process and for dealing with intellectual property.

The detection of trace quantities of aromatic compounds is important to defense and security applications, including the detection of CB agents, explosives, and other substances. These pose threats to forces and the environment. This paper explores an approach to the detection and identification of quantities as little as single molecules of explosives. It can in principle provide instant warning. Apertureless near-field scanning optical microscopy (ANSOM) is one of several promising methods for obtaining spatial resolution below the diffraction limit at various wavelengths, including in the terahertz regime. By scattering incident light off the junction between a probe with a sub-wavelength tip and the surface of a sample, spatial resolution on the order of the tip size can be obtained. For terahertz time-domain spectroscopy where the wavelength-limited resolution is ~1 millimeter, this is a significant advantage. In the case of a sufficiently small probe tip and a thin metallic substrate, plasmonic interaction between the tip and sample provides an enhancement of the near-field in the junction. This effect is dramatically enhanced for nanometer-scale metal layers, since surface plasmon states from both sides of the film can contribute to the overall field enhancement. We present preliminary results of THz plasmonic field enhancements, using a thin (500 nm) gold film evaporated on glass. We observe an enhancement in the scattered THz wave, which we attribute to the large density of plasmonic states extending throughout the THz range. This result indicates a route to single-molecule spectroscopy at terahertz frequencies.

This paper presents significant improvements to our previous watermarking technique for Intellectual Property Protection (IPP) of IP cores. The technique relies on hosting the bits of a digital signature at the HDL design level using resources included within the original system. Thus, any attack trying to change or remove the digital signature will damage the design. The technique also includes a procedure for secure signature extraction requiring minimal modifications to the system. The new advances refer to increasing the applicability of this watermarking technique to any design, not only to those including look-ups, and the provision of an automatic tool for signature hosting purposes. Synthesis results show that the application of the proposed watermarking strategy results in negligible degradation of system performance and very low area penalties and that the use of the automated tool, in addition to easy the signature hosting, leads to reduced area penalties.

An algorithm of dynamic watermark based on neural network is presented which is more robust against attack of false authentication and watermark-tampered operations contrasting with one watermark embedded method. (1) Five binary images used as watermarks are coded into a binary array. The total number of 0s and 1s is 5*N, every 0 or 1 is enlarged fivefold by information-enlarged technique. N is the original total number of the watermarks' binary bits. (2) Choose the seed image pixel p_x,y and its 3×3 vicinities pixel p _x-1,y-1,p_x-1,y,p_x-1,y+1,p_x,y-1,p_x,y+1,p_x+1,y-1,p_x+1,y,p_x+1,y+1 as one sample space. The p_x,y is used as the neural network target and the other eight pixel values are used as neural network inputs. (3) To make the neural network learn the sample space, 5*N pixel values and their closely relevant pixel values are randomly chosen with a password from a color BMP format image and used to train the neural network.(4) A four-layer neural network is constructed to describe the nonlinear mapped relationship between inputs and outputs. (5) One bit from the array is embedded by adjusting the polarity between a chosen pixel value and the output value of the model. (6) One randomizer generates a number to ascertain the counts of watermarks for retrieving. The randomly ascertained watermarks can be retrieved by using the restored neural network outputs value, the corresponding image pixels value, and the restore function without knowing the original image and watermarks (The restored coded-watermark_bit=1, if ox,y(restored)>p_x,y(reconstructed, else coded-watermark_bit =0). The retrieved watermarks are different when extracting each time. The proposed technique can offer more watermarking proofs than one watermark embedded algorithm. Experimental results show that the proposed technique is very robust against some image processing operations and JPEG lossy compression. Therefore, the algorithm can be used to protect the copyright of one important image.

An ultra-wideband (UWB) inter-radio ranging technology with measurement resolution of +/-0.5 ft and range up to 0.5 kilometer under certain FCC regulation was recently introduced. However, measurement data are extremely erroneous due to stochastic variables in the device and multipath radio wave reflections. This paper presents fuzzy logic tuned double tracking filters as a solution to remove misinformation in the data. The 1st tracker locates the overall center of the data in the presence of the large sporadic noise. A fuzzy logic admits only neighborhood data to a 2nd tracker which takes care of smaller deviation noise. The fuzzy neighborhood filter approach has been successfully applied to clean up the UWB radio ranges. Experimental results are shown.

Feature extraction methods based on the statistical analysis of the change in event pressure levels over a period and the level of ambient pressure excitation facilitate the development of a robust classification algorithm. The features reliably discriminates mortar and artillery variants via acoustic signals produced during the launch events. Utilizing acoustic sensors to exploit the sound waveform generated from the blast for the identification of mortar and artillery variants as type A, etcetera through analysis of the waveform. Distinct characteristics arise within the different mortar/artillery variants because varying HE mortar payloads and related charges emphasize varying size events at launch. The waveform holds various harmonic properties distinct to a given mortar/artillery variant that through advanced signal processing and data mining techniques can employed to classify a given type. The skewness and other statistical processing techniques are used to extract the predominant components from the acoustic signatures at ranges exceeding 3000m. Exploiting these techniques will help develop a feature set highly independent of range, providing discrimination based on acoustic elements of the blast wave. Highly reliable discrimination will be achieved with a feedforward neural network classifier trained on a feature space derived from the distribution of statistical coefficients, frequency spectrum, and higher frequency details found within different energy bands. The processes that are described herein extend current technologies, which emphasis acoustic sensor systems to provide such situational awareness.

Pulse Coupled Neural Networks (PCNNs) have been shown to be of value in image processing applications, especially at identifying features of small spatial extent at low signal to noise ratio. In our use of the PCNN, every pixel in a scene feeds a neuron in a fully connected lateral neural network. Nearest neighbor neurons contribute to the output of any given neuron using weights that link the neuron and its neighborhood in both a linear and a non-linear fashion. The network is pulsed, and the output of the network at each pulse is a binary mask of neurons that are active. Pulsing drives the network to evaluate its state. The multi-dimensionality and the non-linear nature of the network make selecting weights using trial and error a non-trivial problem. It is important that the desired features of the input are identified on a predictable pulse, a problem that has yet to be sufficiently addressed by proponents of the PCNN. Our method to overcome these problems is to use a Genetic Algorithm to select the set of PCNN coefficients which will identify the pixels of interest on a predetermined pulse. This method enables PCNNs to be trained, which is a novel capability and renders the method of use for applications.

Airborne LiDAR data are useful for 3D terrain visualization. Segmentation of the data set is especially important for extracting vector data in scenes containing man-made structures. In a graph theoretic formulation, each pixel is a node in a connected graph. The likelihood of an edge existing between two pixels is encoded as the weight between the nodes. Segmentation becomes a graph partitioning that minimizes the weights of the cut links. We combine texture analysis, morphological operators, and the normalized-cut graph-theoretic algorithm to segment lidar data sets. Experimental results using collected data demonstrate the efficacy of our method.

Hard thresholding seems to work well for denoising signals using higher-order statistics. We statistically examined the best values for hard thresholding and related this to the fraction of wavelet coefficients set to zero to obtain the minimum MSE. In addition, we found that the minimum MSE obtained was less sensitive to the threshold when implemented based on a third-order parameter rather than the noise power. Alternatively, we found that this approach to thresholding could be implemented by setting a fixed fraction of wavelet coefficients to zero.

In this paper we introduce a canonical minimised adder graph (CMAG) representation that can easily be generated with a computer. We show that this representation can be used to efficiently develop code generation for MAG graphs. Several code optimizations methods are developed in the computation of the non-output fundamental sum (NOFS) computation, which allows the computation of all graphs up to cost-5 be accomplished in a reasonable timeframe.

Knowledge-based clustering and autonomous mental development remains a high priority research topic, among which the learning techniques of neural networks are used to achieve optimal performance. In this paper, we present a new framework that can automatically generate a relevance map from sensory data that can represent knowledge regarding objects and infer new knowledge about novel objects. The proposed model is based on understating of the visual what pathway in our brain. A stereo saliency map model can selectively decide salient object areas by additionally considering local symmetry feature. The incremental object perception model makes clusters for the construction of an ontology map in the color and form domains in order to perceive an arbitrary object, which is implemented by the growing fuzzy topology adaptive resonant theory (GFTART) network. Log-polar transformed color and form features for a selected object are used as inputs of the GFTART. The clustered information is relevant to describe specific objects, and the proposed model can automatically infer an unknown object by using the learned information. Experimental results with real data have demonstrated the validity of this approach.

Breast cancer has been one of the leading causes of cancer deaths for females in the developed countries, including the US. While early detection of breast cancer is essential for the reduction of death rate, there may be already more than 10⁷ cells in a breast cancer when it can be observed by X-ray mammogram. As contrast, the passive IR spectrogram proposed by Szu et al. was shown to be promising in detecting the breast cancer several months ahead of mammogram. With the energy readings from two IR cameras, one middle wavelength IR (MIR, 3 - 5μm) and one long wavelength IR (LIR, 8 - 12μm), the IR spectrogram may be computed by using the blind source separation (BSS) algorithms developed by Szu et al., which reveals the probability of being a cancer point on the breast surface. Two important tasks are involved in computing the IR spectrogram. One is an accurate estimate of the ground state energy in the Helmholtz free energy, H = E-T₀S. The other is a correct pair-up of the points on the MIR and LIR images for a better estimation of IR spectrogram. To minimize the probability of making an erroneous estimate of the ground state energy inherent in the deterministic neighborhood-based BSS algorithm, a spatiotemporal approach is proposed in this paper. It takes into account not only the neighborhood information but also the temporal information in determining the probability of being a cancer point. Furthermore, a new sub-pixel super-resolution registration algorithm incorporating a third energy dimension is proposed to establish better correspondences between the points in the MIR and LIR images. Phantom study has confirmed that sub-pixel registration can be achieved by the proposed registration method. Human subject study further shows that the breast cancer may be detected by the proposed spatiotemporal approach via cross-referencing the IR spectrograms computed from the multiple pairs of MIR and LIR images taken at different times.

Today robotic technology can be developed and advanced a lot from industrial robotic manipulation to the micro and nano manipulations. We have been proposed "Nano Laboratory" consisted of the capabilities of nanofabrication, nanoinstrumentaion and nanoassembly to make three dimensional structure, devices and systems in the nanoworld. First the nanorobotic manipulation system is introduced to realize these capabilities inside a scanning electron microscope (SEM) and a transmission electron microscope (TEM). Then some precursors are introduced into the working small space in the nanoworld, so that the cutting, bending and fixing operation in the nanoworld will be realized using nanomanipulator, electron-bean-induced deposition (EBID) and other methodologies as in the macro world. After making the three dimensional structure, nano devices such as sensor and actuator can be fabricated.

This paper reviews the DARPA MTO STAP-BOY program for both Phase I and II. The STAP-BOY program conducts fast covariance factorization and tuning techniques for space-time adaptive process (STAP) Algorithm Implementation on Graphics Processor unit (GPU) Architectures for Embedded Systems. Emerging capabilities in stream and multi-core computation, along with high speed memory bandwidths in commercial GPU architectures, are enabling breakthrough low-cost and low-power teraflop computing solutions to DoD-embedded computing challenges. Under the DARPA MTO STAP-BOY program, SAIC and Duke University, in cooperation with commercial graphics processor companies, have been mapping complex signal processing algorithms to GPU architectures. Algorithms undergoing implementation include STAP applications for radar adaptive beamforming and spin-image surface matching applications for object recognition in 3-D range-image data.

This paper reviews the implementation of DARPA MTO STAP-BOY program for both Phase I and II conducted at Science Applications International Corporation (SAIC). The STAP-BOY program conducts fast covariance factorization and tuning techniques for space-time adaptive process (STAP) Algorithm Implementation on Graphics Processor unit (GPU) Architectures for Embedded Systems. The first part of our presentation on the DARPA STAP-BOY program will focus on GPU implementation and algorithm innovations for a prototype radar STAP algorithm. The STAP algorithm will be implemented on the GPU, using stream programming (from companies such as PeakStream, ATI Technologies' CTM, and NVIDIA) and traditional graphics APIs. This algorithm will include fast range adaptive STAP weight updates and beamforming applications, each of which has been modified to exploit the parallel nature of graphics architectures.

Accurate image registration (with subpixel accuracy) techniques are critical components for many advanced image processing systems such as the time-differencing system and the super-resolution imaging system. In this paper, several image registration methods are compared and evaluated. Performance of registration accuracy is evaluated using several LWIR and CCD video imagery by two metrics: RMSE (root mean square error) and SCNR (signal-to-clutter noise-ratio) gain (improvement of SCNR). Promising applications for iris biometric identification, super-resolution image enhancements, heavy background clutter suppression for improving moving target detection have been presented in this paper.