Search

E. coli promoter recognition is an area of great interest in bioinformatics. In this paper, we describe the implementation of a feed forward neural network to predict the E. coli promoter. According to the sequence conservation, some sequences with 60 bases are selected as positive samples and some corresponding non-promoters from E. coli coding areas are selected as negative samples, and a classifier based on feed forward neural network is trained. Results show that feed forward neural networks can extract the statistical characteristics of promoters more effectively, and that coding with four dimensions for nucleic acid data is superior to two dimensions. Another result demonstrated here is that the number of hidden layers seems to have no significant effect on E. coli promoter prediction precision. The research results in this paper can provide reference for promoter recognition research.

In this study, we decomposed 12 fMRI data sets from six subjects each 101 times using the infomax algorithm. The first decomposition was taken as a reference decomposition; the others were used to form a component matrix of 100 by 100 components. Equivalence relations between components in this matrix, defined as maximum spatial correlations to the components of the reference decomposition, were found by the Hungarian sorting method and used to form 100 equivalence classes for each data set. We then tested the reproducibility of the matched components in the equivalence classes using uncertainty measures based on component distributions, time courses, and ROC curves. Infomax ICA rarely failed to derive nearly the same components in different decompositions. Very few components per data set were poorly reproduced, even using vector angle uncertainty measures stricter than correlation and detection theory measures.

Current optical coherence tomography (OCT) relies on patient cooperation to maintain fixation. For patients with eye complications, it is extremely difficult to achieve fixation on the eye. Most eye trackers are external and track only the pupils. Such method is not useful for the OCT. The only existing fundus image based tracker is a hardware-based system that is composed of a confocal reflectometer, dither scanner, and tracking galvanometers. The integration of such hardware based system with the OCT is not as desirable nor as portable. Therefore, we want an eye tracking system that can be installed in the computer connected to the OCT. In this paper, a real time internal eye tracking software based system is proposed. By combining intensity based object tracking adaptive mean shift algorithm with Kalman filtering techniques together, a robust real time eye tracking algorithm is proposed. The system is robust in terms of its acceptance to low resolution, coarsely quantized, or extremely noisy images and still able to produce desirable results with no pre-processing of images.

We have developed an integrated dynamic lens system for in vivo optical imaging. Bioinspired dynamic microfluidic lenses allow for real-time dynamic manipulation of the lens focal length via microfluidic injection into a PDMS membrane-capped chamber. A piezoelectrically actuated micropump is integrated with with the lens to provide highspeed, accurate lens tunability. The 5mm dynamic lens has demonstrated focal length tunability from 8.5mm to 23mm, numerical aperture values from 0.39 to 0.77, and resolution of 40 linepairs/mm. The micropump operates at 5 kHz and achieved a flow rate of approximately 2.4 mL/min. This system can be applied to optical probe techniques to improve diagnosis with real-time depth resolution and variable numerical aperture.

A technique of detecting the rolling and spinning of half-coated nanoparticles using interference ring patterns of the fluorescence has been applied to the measurement of torsional spring modulus of a double-stranded DNA. Using the unique ability to measure nanoparticle rotations in multiple degrees of freedom, we were able to determine the spinning of a nanoparticle tethered on the DNA and thereby the twisting of the DNA in real time. The detailed knowledge of the spinning as well as rolling behaviors of half-coated nanoparticles provides information about torsional elastic properties of the DNA under investigation.

An unsupervised classification algorithm is derived by modeling observed data as a mixture of several mutually exclusive classes that are each described by linear combinations of independent, non-Gaussian densities. The algorithm estimates the data density in each class by using parametric nonlinear functions that fit to the non-Gaussian structure of the data. This improves classification accuracy compared with standard Gaussian mixture models. When applied to images, the algorithm can learn efficient codes (basis functions) for images that capture the statistically significant structure intrinsic in the images. We apply this technique to the problem of unsupervised classification, segmentation, and denoising of images. We demonstrate that this method was effective in classifying complex image textures such as natural scenes and text. It was also useful for denoising and filling in missing pixels in images with complex structures. The advantage of this model is that image codes can be learned with increasing numbers of classes thus providing greater flexibility in modeling structure and in finding more image features than in either Gaussian mixture models or standard independent component analysis (ICA) algorithms.

A number of current face recognition algorithms use face representations found by unsupervised statistical methods. Typically these methods find a set of basis images and represent faces as a linear combination of those images. Principal component analysis (PCA) is a popular example of such methods. The basis images found by PCA depend only on pairwise relationships between pixels in the image database. In a task such as face recognition, in which important information may be contained in the high-order relationships among pixels, it seems reasonable to expect that better basis images may be found by methods sensitive to these high-order statistics. Independent component analysis (ICA), a generalization of PCA, is one such method. We used a version of ICA derived from the principle of optimal information transfer through sigmoidal neurons. ICA was performed on face images in the FERET database under two different architectures, one which treated the images as random variables and the pixels as outcomes, and a second which treated the pixels as random variables and the images as outcomes. The first architecture found spatially local basis images for the faces. The second architecture produced a factorial face code. Both ICA representations were superior to representations based on PCA for recognizing faces across days and changes in expression. A classifier that combined the two ICA representations gave the best performance.

One method of transmitting wavelet based zerotree encoded images over noisy channels is to add channel coding without altering the source coder. A second method is to reorder the embedded zerotree bitstream into packets containing a small set of wavelet coefficient trees. We consider a hybrid mixture of these two approaches and demonstrate situations in which the hybrid image coder can outperform either of the two building block methods, namely on channels that can suffer packet losses as well as statistically varying bit errors.

This paper demonstrates the use of a fiberoptic bundle for flexible, compact, remote, and noncontact laser generation of surface ultrasonic waves in materials. The bundle is able to deliver Nd:YAG pulsed light with a 60% delivery efficiency up to an average energy of 55 mJ/pulse for a pulse duration on the order of 10 ns and a pulse repetition rate of 20 Hz without signs of fiber damage. Details of the bundle construction and surface preparation are given, and pulsed light delivery tests performed with single tapered fibers are discussed. The high-power light delivery capabilities of the bundle are demonstrated for the generation of narrow-band surface waves in a Carbon/PEEK composite laminate by a spatial modulation technique that employs a periodic transmission mask. Single laser pulse ultrasonic tonebursts are clearly detectable using a small aperture piezoelectric transducer while ensuring thermoelastic generation conditions. The theory of narrow-band generation of surface acoustic waves is improved by accounting for the strength nonuniformity of the illumination sources. In addition, the effect of the number of illumination sources on the bandwidth of the generated surface wave is assessed experimentally, and excellent agreement is shown with the theoretical results predicted by the improved model.

For the last 10 years, interest has grown in low frequency shear waves that propagate in the human body. However, the generation of shear waves by acoustic vibrators is a relatively complex problem, and the directivity patterns of shear waves produced by the usual vibrators are more complicated than those obtained for longitudinal ultrasonic transducers. To extract shear modulus parameters from the shear wave propagation in soft tissues, it is important to understand and to optimize the directivity pattern of shear wave vibrators. This paper is devoted to a careful study of the theoretical and the experimental directivity pattern produced by a point source in soft tissues. Both theoretical and experimental measurements show that the directivity pattern of a point source vibrator presents two very strong lobes for an angle around 35 degrees . This paper also points out the impact of the near field in the problem of shear wave generation.

A Neural integrated Fuzzy conTroller (NiF-T) which integrates the fuzzy logic representation of human knowledge with the learning capability of neural networks is developed for nonlinear dynamic control problems. NiF-T architecture comprises of three distinct parts: (1) Fuzzy logic Membership Functions (FMF), (2) a Rule Neural Network (RNN), and (3) an Output-Refinement Neural Network (ORNN). FMF are utilized to fuzzify sensory inputs. RNN interpolates the fuzzy rule set; after defuzzification, the output is used to train ORNN. The weights of the ORNN can be adjusted on-line to fine-tune the controller. In this paper, real-time implementations of autonomous mobile robot navigation and multirobot convoying behavior utilizing the NiF-T are presented. Only five rules were used to train the wall following behavior, while nine were used for the hall centering. Also, a robot convoying behavior was realized with only nine rules. For all of the described behaviors-wall following, hall centering, and convoying, their RNN's are trained only for a few hundred iterations and so are their ORNN's trained for only less than one hundred iterations to learn their parent rule sets.

In this work, we characterize and contrast the capabilities of the general class of time-delay neural networks (TDNNs) with input delay neural networks (IDNNs), the subclass of TDNNs with delays limited to the inputs. Each class of networks is capable of representing the same set of languages, those embodied by the definite memory machines (DMMs), a subclass of finite-state machines. We demonstrate the close affinity between TDNNs and DMM languages by learning a very large DMM (2048 states) using only a few training examples. Even though both architectures are capable of representing the same class of languages, they have distinguishable learning biases. Intuition suggests that general TDNNs which include delays in hidden layers should perform well, compared to IDNNs, on problems in which the output can be expressed as a function on narrow input windows which repeat in time. On the other hand, these general TDNNs should perform poorly when the input windows are wide, or there is little repetition. We confirm these hypotheses via a set of simulations and statistical analysis.

The problem of sensorimotor control is underdetermined due to excess (or "redundant") degrees of freedom when there are more joint variables than the minimum needed for positioning an end-effector. A method is presented for solving the nonlinear inverse kinematics problem for a redundant manipulator by learning a natural parameterization of the inverse solution manifolds with self-organizing maps. The parameterization approximates the topological structure of the joint space, which is that of a fiber bundle. The fibers represent the "self-motion manifolds" along which the manipulator can change configuration while keeping the end-effector at a fixed location. The method is demonstrated for the case of the redundant planar manipulator. Data samples along the self-motion manifolds are selected from a large set of measured input-output data. This is done by taking points in the joint space corresponding to end-effector locations near "query points", which define small neighborhoods in the end-effector work space. Self-organizing maps are used to construct an approximate parameterization of each manifold which is consistent for all of the query points. The resulting parameterization is used to augment the overall kinematics map so that it is locally invertible. Joint-angle and end-effector position data, along with the learned parameterizations, are used to train neural networks to approximate direct inverse functions.

Many of the properties of the well-known Kohonen map algorithm are not easily derivable from its discrete formulation. For instance, the "projection" implemented by the map from a high dimensional input space to a lower dimensional map space must be properly regarded as a projection from a smooth manifold to a lattice and, in this framework, some of its properties are not easily identified. This paper describes the self-organizing field: a continuous embedding of a smooth manifold (the map) into another (the input manifold) that implements a topological map by self-organization. The adaptation of the self-organizing field is governed by a set of differential equations analogous to the difference equations that determine weights updates in the Kohonen map. This paper derives several properties of the self-organizing field, and shows that the emergence of certain structures on the brain-like the columnar organization in the primary visual cortex-arise naturally in the new model.

Subband and wavelet decompositions are powerful tools in image coding because of their decorrelating effects on image pixels, the concentration of energy in a few coefficients, their multirate/multiresolution framework, and their frequency splitting, which allows for efficient coding matched to the statistics of each frequency band and to the characteristics of the human visual system. Vector quantization (VQ) provides a means of converting the decomposed signal into bits in a manner that takes advantage of remaining inter and intraband correlation as well as of the more flexible partitions of higher dimensional vector spaces. Since 1988, a growing body of research has examined the use of VQ for subband/wavelet transform coefficients. We present a survey of these methods.

Localizing boundaries between textured image regions without sacrificing the labeling accuracy of interior regions remains a problem in segmentation. Difficulties arise because of the conflicting requirements of localization and labeling. Boundary localization usually demands observing the features over small neighborhoods, whereas labeling accuracy increases with the size of the observation neighborhood. This problem is further exacerbated in texture segmentation by the spatially distributed nature of texture features. In this correspondence, we develop a multiresolution approach that combines localized and distributed features to directly address boundary localization in texture segmentation. Maximum localization is achieved by using the gray-level discontinuities at the boundary between textures to define the boundary. The properties that characterize the gray-level discontinuity at texture boundaries are developed and an algorithm is designed to localize the boundary using these discontinuities. This segmentation algorithm is implemented and successfully tested on a set of Brodatz texture mosaics and AVHRR satellite imagery.

A radial basis function neural network was successfully applied to an area which is relatively new for neural networks: a remote sensing application that provides estimates of water vapor content, an important parameter for climate modeling. The neural network provided results which are up to 32% better than had been previously obtained using conventional statistical methods on the same data. These results have implications for improved short-term weather forecasting and for long-term global climate modeling. The neural network approach is compared with the past and present operating algorithms at the National Oceanic and Atmospheric Administration. The radial basis function network's performance is compared with sigmoidal backpropagation network. Low-power electronic implementations of the neural methodology were explored to demonstrate the feasibility of placing the network on a remote sensing platform. This would permit processing the raw sensor data into information on the platform, eliminating the need to store the raw data, and helping to contain the expected explosion of climate data.

A supervised learning neural network (SLNN) coprocessor which enhances the performance of a digital soft-decision Viterbi decoder used for forward error correction in a digital communication channel with either fading plus additive white Gaussian noise (AWGN) or pure AWGN has been investigated and designed. The SLNN is designed to cooperate with a phase shift keying (PSK) demodulator, an automatic gain control (AGC) circuit, and a 3-bit quantizer which is an analog to digital convertor. It is trained to learn the best uniform quantization step-size Delta (BEST) as a function of the mean and the standard deviation of various sets of Gaussian distributed random variables. The channel cutoff rate (R(0)) of the channel is employed to determine the best quantization threshold step-size (Delta(BEST)) that results in the minimization of the Viterbi decoder output bit error rate (BER). For a digital communication system with a SLNN coprocessor, consistent and substantial BER performance improvements are observed. The performance improvement ranges from a minimum of 9% to a maximum of 25% for a pure AWGN channel and from a minimum of 25% to a maximum of 70% for a fading channel. This neural network coprocessor approach can be generalized and applied to any digital signal processing system to decrease the performance losses associated with quantization and/or signal instability.

Probability density functions are estimated by an exponential family of densities based on multilayer feedforward networks. The role of the multilayer feedforward networks, in the proposed estimator, is to approximate the logarithm of the probability density functions. The method of maximum likelihood is used, as the main contribution, to derive an unsupervised backpropagation learning law to estimate the probability density functions. Computer simulation results demonstrating the use of the derived learning law are presented.

This paper discusses the use of evolutionary programming (EP) for computer-aided design and testing of neural controllers applied to problems in which the system to be controlled is highly uncertain. Examples include closed-loop control of drug infusion and integrated control of HVAC/lighting/utility systems in large multi-use buildings. The method is described in detail and applied to a modified Cerebellar Model Arithmetic Computer (CMAC) neural network regulator for systems with unknown time delays. The design and testing problem is viewed as a game, in that the controller is chosen with a minimax criterion i.e., minimize the loss associated with its use on the worst possible plant. The technique permits analysis of neural strategies against a set of feasible plants. This yields both the best choice of control parameters and identification of that plant which is most difficult for the best controller to handle.

This work studies the performance of dimensional least mean square (TDLMS) adaptive filters as prewhitening filters for the detection of small objects in image data. The object of interest is assumed to have a very small spatial spread and is obscured by correlated clutter of much larger spatial extent. The correlated clutter is predicted and subtracted from the input signal, leaving components of the spatially small signal in the residual output. The receiver operating characteristics of a detection system augmented by a TDLMS prewhitening filter are plotted using Monte-Carlo techniques. It is shown that such a detector has better operating characteristics than a conventional matched filter in the presence of correlated clutter. For very low signal-to-background ratios, TDLMS-based detection systems show a considerable reduction in the number of false alarms. The output energy in both the residual and prediction channels of such filters is shown to be dependent on the correlation length of the various components in the input signal. False alarm reduction and detection gains obtained by using this detection scheme on thermal infrared sensor data with known object positions is presented.

The authors derive a method for selecting exemplars for training a multilayer feedforward network architecture to estimate an unknown (deterministic) mapping from clean data, i.e., data measured either without error or with negligible error. The objective is to minimize the data requirement of learning. The authors choose a criterion for selecting training examples that works well in conjunction with the criterion used for learning, here, least squares. They proceed sequentially, selecting an example that, when added to the previous set of training examples and learned, maximizes the decrement of network squared error over the input space. When dealing with clean data and deterministic relationships, concise training sets that minimize the integrated squared bias (ISB) are desired. The ISB is used to derive a selection criterion for evaluating individual training examples, the DISB, that is maximized to select new exemplars. They conclude with graphical illustrations of the method, and demonstrate its use during network training. Experimental results indicate that training upon exemplars selected in this fashion can save computation in general purpose use as well.

The design of a scalable, fully connected 3-D optoelectronic neural system that uses free-space optical interconnects with silicon-VLSI-based hybrid optoelectronic circuits is proposed. The system design uses a hardware-efficient combination of pulsewidth-modulating optoelectronic neurons and pulse-amplitude-modulating electronic synapses. Low-area, high-linear-dynamic-range analog synapse and neuron circuits are proposed. SPICE circuit simulations and an experimental demonstration of the free-space optical interconnection system are included.

Neural network simulations on a parallel architecture are reported. The architecture is scalable and flexible enough to be useful for simulating various kinds of networks and paradigms. The computing device is based on an existing coarse-grain parallel framework (INMOS transputers), improved with finer-grain parallel abilities through VLSI chips, and is called the Lneuro 1.0 (for LEP neuromimetic) circuit. The modular architecture of the circuit makes it possible to build various kinds of boards to match the expected range of applications or to increase the power of the system by adding more hardware. The resulting machine remains reconfigurable to accommodate a specific problem to some extent. A small-scale machine has been realized using 16 Lneuros, to experimentally test the behavior of this architecture. Results are presented on an integer version of Kohonen feature maps. The speedup factor increases regularly with the number of clusters involved (to a factor of 80). Some ways to improve this family of neural network simulation machines are also investigated.

Ferroelectric lead lanthanum zirconate titanate (PLZT) films are deposited on R-plane sapphire using RF triode magnetron sputtering. Perovskite PLZT films with the desired composition (9/65/35) are obtained using compensated deposition techniques around 500 degrees C and postdeposition annealing at 650 degrees C. The deposited films exhibit good optical and electrooptical properties. The room temperature dielectric constant of the films was 1800 at 10 kHz. The refractive index of the films was in the range of 2.2-2.5. The films showed a quadratic electrooptic effect with R=0.6 x10(-16) m(2)/V(2). The development of PLZT on silicon-on-sapphire smart spatial light modulators using these films is also explored.

Blood vessels usually have poor local contrast, and the application of existing edge detection algorithms yield results which are not satisfactory. An operator for feature extraction based on the optical and spatial properties of objects to be recognized is introduced. The gray-level profile of the cross section of a blood vessel is approximated by a Gaussian-shaped curve. The concept of matched filter detection of signals is used to detect piecewise linear segments of blood vessels in these images. Twelve different templates that are used to search for vessel segments along all possible directions are constructed. Various issues related to the implementation of these matched filters are discussed. The results are compared to those obtained with other methods.

The fractal dimension (D(F)) is one measure of the space-filling features of a self-similar structure. Additionally, since D(F) varies with branching level, there may be potential critical locations that are functionality important. The authors introduce an algorithm that models lung airway structures and uses computer simulations of growth based on fractal concepts. Under these conditions, limits imposed by simple boundary constraints generate structures that are in good agreement with actual morphometric data.