From the Publisher: The main thrust is to provide students with a solid understanding of number important and related advanced topics in digital signal processing such as Wiener filters, power spectrum estimation, modeling adaptive filtering. Scores worked examples illustrate fine points, compare techniques algorithms facilitate comprehension fundamental concepts. Also features an abundance interesting challenging problems at end every chapter.

This paper addresses two fundamental issues involved in the reconstruction of a multidimensional sequence from either phase or magnitude its Fourier transform. The first issue relates to uniqueness terms magnitude. Although information alone is not sufficient, general, uniquely specify sequence, large class sequences are shown be recoverable their second which addressed this concerns actual For those specified by phase, several practical algorithms described may used reconstruct phase....

In this paper, we develop a set of conditions under which sequence is uniquely specified by the phase or samples its Fourier transform, and similar magnitude transform. These are distinctly different from minimum maximum conditions, applicable to both one-dimensional multidimensional sequences. Under also several algorithms may be used reconstruct magnitude.

A new night-time lane detection system and its accompanying framework are presented in this paper. The consists of an automated ground truth process systematic storage captured videos that will be used for training testing. proposed Advanced Lane Detector 2.0 (ALD 2.0) is improvement over the ALD 1.0 or Layered Approach with integration pixel remapping, outlier removal, prediction tracking. Additionally, a novel procedure to generate data marker locations also proposed. original called time...

Face images that are captured by surveillance cameras usually have a very low resolution, which significantly limits the performance of face recognition systems. In past, super-resolution techniques been proposed to increase resolution combining information from multiple images. These use as preprocessing step obtain high-resolution image is later passed system. Considering most state-of-the-art systems an initial dimensionality reduction method, we propose transfer reconstruction pixel...

The work presented in this paper focuses on the use of hidden Markov models for face recognition. A new method based extraction 2D-DCT feature vectors is described, and recognition results are compared with other approaches. introduced reduces significantly computational complexity previous HMM-based system, while preserving same rate.

Our goal is to exploit human motion and object context perform action recognition classification. Towards this end, we introduce a framework for recognizing actions objects by measuring image-, object- action-based information from video. Hidden Markov models are combined with classify hand actions, which aggregated Bayesian classifier summarize activities. We also use methods differentiate the class of unknown evaluating detected along low-level, extracted features. approach appropriate...

In a previous paper, simple approach to lane detection using the Hough transform and iterated matched filters was described. This paper extends this work by incorporating an inverse perspective mapping create bird's-eye view of road, applying random sample consensus help eliminate outliers due noise artifacts in Kalman filter smooth output tracker.

Two iterated function system (IFS) models are explored for the representation of single-valued discrete-time sequences: self-affine fractal model and piecewise model. Algorithms presented, one which is suitable a multiprocessor implementation, identification parameters each Applications these to variety data types given where signal-to-noise ratios quantization effects investigated, compression computed.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML"...

Abstract NOTE: The first page of text has been automatically extracted and included below in lieu an abstract Session 2532 Distance Learning Into the 21st Century Monson H. Hayes Lonnie Harvel School Electrical Computer Engineering Georgia Institute Technology Atlanta, GA 30332 is a unique position having campus located across Atlantic heart Lorraine region France. Offering masters degree programs electrical mechanical engineering, Tech (GTL) must deliver more courses than can be staffed by...

We describe an embedded hidden Markov model (HMM)-based approach for face detection and recognition that uses efficient set of observation vectors obtained from the 2D-DCT coefficients. The HMM can two dimensional data better than one-dimensional is computationally less complex two-dimensional HMM. This appropriate images since it exploits important facial characteristic: frontal faces preserve same structure "super states" top to bottom, also left-to-right "states" inside each these states".

The work presented in this paper describes a hidden Markov model (HMM)-based framework for face recognition and detection. observation vectors used to characterize the states of HMM are obtained using coefficients Karhunen-Loeve transform (KLT). method reduces significantly computational complexity previous HMM-based systems, while slightly improving rate. Consistent with face, introduces novel detection approach same feature extraction techniques recognition.

Polynomials in more than one variable arise frequently multidimensional signal processing applications. Unlike polynomials a single variable, cannot, general, be factored. In this note, it is shown that the set of factorable multi-dimensional extremely small sense almost all two or variables are irreducible.

Iterative algorithms for signal reconstruction from partial time- and frequency-domain knowledge have proven useful in a number of application areas. In this paper, general convergence proof, applicable to class such iterative algorithms, is presented. The proof relies on the concept nonexpansive mapping both time frequency domains. Two examples studied detail are time-limited extrapolation (equivalently, band-limited extrapolation) phase-only reconstruction. iteration new result obtained by...

A layered approach is designed to address many of the real-world problems that an inexpensive lane detection system would encounter. region interest first extracted from image followed by enhancement procedure manipulate shape markers. The then converted binary using adaptive threshold. model based line hypothesizes position. Finally, iterated matched filtering scheme estimates final developed shows good performance when tested on data contains fluctuating illumination and a variety traffic...

The embedded hidden Markov model (HMM) is a statistical that can be used in many pattern recognition and computer vision applications. This inherits the partial size invariance of standard HMM, and, due to its pseudo two-dimensional structure, able data such as images, better than HMM. We describe maximum likelihood training for continuous mixture HMM present performance this face detection recognition. experimental results are compared with other approaches

The active appearance model (AAM) is a powerful tool for modeling images of deformable objects and has been successfully used in variety alignment, tracking, recognition applications. AAM uses subspace-based models to represent the certain object class. In general, fitting such complicated previously unseen using standard optimization techniques computationally complex task because gradient matrix be numerically computed at every iteration. critical feature fast convergence scheme which...

In this paper, we show that a one-dimensional or multidimensional sequence is uniquely specified under mild restrictions by its signed Fourier transform magnitude (magnitude and 1 bit of phase information). addition, develop numerical algorithm to reconstruct from magnitude. Reconstruction examples obtained using are also provided.

Consumer video surveillance systems often suffer from bad weather conditions, observed objects lose visibility and contrast due to the presence of atmospheric haze, fog, smoke. In this paper, we present an image defogging algorithm with color correction in HSV space for processing. We first generate a modified transmission map segmentation using multiphase level set formulation intensity (V) values. also estimate light The proposed method can significantly enhance foggy frames estimated map....

In communications, innovative paradigm shifts have emerged in integrating various devices into the network to provide advanced and intelligent services. However, security threats may occur that not always be detected using traditional cryptographic techniques. Secure authentication is of paramount importance modern wireless systems. This paper focusses on robust a time-varying communication environment where conventional mechanisms are severely limited. We propose an Adaptive Neural Network...

In this paper, we present a segmented linear subspace model for face recognition that is robust under varying illumination conditions. The algorithm generalizes the 3D by segmenting image into regions have surface normals whose directions are close to each other. This segmentation performed using K-means clustering and requires only few training images different illuminations. When applied image, attached cast shadows, rate equal of computationally more complex systems require constructing face.

Wireless networks strive to integrate information technology into every corner of the world. This openness radio propagation is one reason why holistic wireless security mechanisms only rarely enter picture. In this paper, we propose a physical (PHY)-layer authentication scheme that takes advantage channel randomness detect spoofing attacks in networks. Unlike most existing techniques rely on comparing message between legitimate user and potential spoofer, our proposed uses Gaussian mixture...

In this paper, we present a novel high dynamic range (HDR) imaging method using single image. The existing multiple image-based HDR methods work only on condition that there is no camera and object movement when acquiring multiple, differently exposed low (LDR) images. To overcome such an unrealistic restriction, make three LDR images from input image local histogram stretching. An edge-preserving spatially adaptive denoising also proposed to suppress the noise amplified in stretching...

In this paper, we propose a new methodology for detecting lane markers that exploits the parallel nature of boundaries on road. First, input image is pre-processed and Altered to detect marker features. Then, using Polar Randomized Hough Transform introduced in lines are fitted through detected features orientation each line evaluated. By finding near separated by constraint specified distance, false signalling caused artifacts greatly reduced. The proposed system was tested real world...