Alan J. Van Nevel

Comparative study of maximum average correlation height filter variants using ladar imagery

We present a study of a family of maximum average correlation height (MACH) filters. MACH filters were introduced by Mahalanobis et al., and several modifications such as the extended MACH (EMACH) and generalized MACH (GMACH) have been introduced to enhance the utility of the MACH filter approach. A comparison between the different filtering approaches and processing techniques is presented for the specific case of laser radar (ladar) imagery. The comparison utilizes both synthetic data for training and testing in one case, and synthetic training data and collected ladar imagery for testing in the second case. The results indicate that the GMACH variant of MACH filters is superior to both MACH and EMACH filters for the case of laser radar.

Signal-to-clutter measure for measuring automatic target recognition performance using complimentary eigenvalue distribution analysis

The performance of an automatic target recognition (ATR) algorithm is not only influenced by the relevance of the prior (training) information, but also by the level of difficulty posed by the clutter that surrounds the target. Thus, an objective measure of target- (or signal-) to-clutter ratio (SCR) is important for the assessment of ATRs. We describe a new metric for SCR based on an eigen analysis of a two-class problem. It is believed that an SCR metric, along with a measure for the relevancy of the training data, are the key parameters for the characterization of ATR performance. Various examples are given to illustrate the application of eigen analysis in determining the difficulty level of finding a particular target in the presence of clutter, and consequently how this new SCR metric defines the potential for false alarms.

Integrated approach for automatic target recognition using a network of collaborative sensors

We introduce what is believed to be a novel concept by which several sensors with automatic target recognition (ATR) capability collaborate to recognize objects. Such an approach would be suitable for netted systems in which the sensors and platforms can coordinate to optimize end-to-end performance. We use correlation filtering techniques to facilitate the development of the concept, although other ATR algorithms may be easily substituted. Essentially, a self-configuring geometry of netted platforms is proposed that positions the sensors optimally with respect to each other, and takes into account the interactions among the sensor, the recognition algorithms, and the classes of the objects to be recognized. We show how such a paradigm optimizes overall performance, and illustrate the collaborative ATR scheme for recognizing targets in synthetic aperture radar imagery by using viewing position as a sensor parameter.

A collaborative network of correlation filters for object recognition

In this paper, we introduce a novel method where several sensors and ATRs collaborate to recognize objects. Such an approach would be suitable for network centric application where the sensors and platforms can coordinate to optimize over all ATR performance. We use correlation pattern recognition techniques to facilitate the development of the concept, although other algorithms may be easily substituted. Essetnially, a self-configuring network is proposed that positions the sensors optimally with respect to each other depending on the algorithm and the class of the object to be recognized. We show how such a network optimizes overall performance, and illustrate the scheme by means of examples.

Locating Peaks in Proteomic Mass Spectral Data Using the Morel-Helmholtz Principle

Proteomics is a rapidly emerging field of research that will help identify and characterize the complex proteins that are responsible for the function of complex biological systems. For detection and identification of separated components, mass spectrometry is evolving to be the method-of-choice because of its high sensitivity and its ability to characterize the individual components. Analysis of biological sample will typically generate a protein mass fingerprint of the various constitutive components, with the component mass expressed as mass-to-charge (m/z) ratios and the relative abundance of each component as the peak height. However, reliably finding protein peaks with small relative abundance has been a difficult signal processing task, and many of the currently used techniques require many arbitrary parameters. This paper investigates the application of the Morel-Helmholtz principle, a single parameter method, to mass spectrometry signal processing. A comparison of the Morel-Helmholtz peak finding method with a thresholding method demonstrates that using the false alarm rate of one per interval will detect peaks that can optimally classify mass spectrometry data equally well as a well chosen threshold

Volume correlation filters for recognizing patterns in 3D data

Correlation filters are ideally suited for recognizing patterns in three-dimensional (3D) data. Whereas most model-based techniques tend to measure the overall dimensions of objects and their larger features, correlation filters can readily (and efficiently) exploit intricate surface details, the gray values of surfaces as well as internal structure, if any. Thus correlation filters may be the preferred approach in scenarios when intensity and range data are both available, or when the internal structure of an object has been mapped (e.g. tomography). In this paper, we outline the development of filters for 3D data that we refer to as Volume Correlation Filters (VCFs), illustrate their use with range images of an object, and outline future work for the development of 3D correlation techniques.

Geometric diffusion as a classifier

Often, the problem of automatic target recognition can be reduced down to two separate but related problems, feature extraction and classifier design. The best classifier only works as well as the input data provided to the system. In this presentation, we will outline a new approach to classification known as geometric diffusion as proposed by Coifman et al, and demonstrate the power of this new metric for classification of imagery.

Automatic registration of sequential ladar imagery

In this work we present an algorithm used to automatically register a sequence of ladar images taken from a sensor flying a known flight path . The registration is performed with no human operator intervention. The resulting mosaic is accurate to one pixel. The algorithmic approach was developed in such a way to allow for near real time processing. The initial method can be extended to registration of multiple views of a scene, with four degrees of freedom (translation and in plane rotation). In this work we will restrict ourselves to rigid body transformations. The registered mosaic is an important step in geolocation using a reference digital elevation map, and will be explored in future work.

Three-dimensional correlation filters for orientation invariant recognition

Correlation filters are ideally suited for recognizing patterns in 3D data. Whereas most model-based techniques tend to measure the overall dimensions of objects and their larger features, correlation filters can readily exploit intricate surface details, the gray values of surfaces as well as internal structure, if any. Thus correlation filters may be the preferred approach in scenarios when intensity and range data are both available, or when the internal structure of an object has been mapped. In this paper, we outline the development of filters for 3D data that we refer to as Volume Correlation Filters, illustrate their use with range images of an object, and outline future work for the development of 3D correlation techniques.