Multispectral Image Processing and Pattern Recognition

The following sections are included:

• CONTENTS

The papers of this book were selected from those presented at the First International Symposium on Multispectral Image Processing And Pattern Recognition, co-sponsored by SPIE, held on October 21–23, 1998 in Wuhan, China…

This paper is aimed at 3D object understanding from 2D images, including articulated objects in active vision environment, using interactive, and internet virtual reality techniques. Generally speaking, an articulated object can be divided into two portions: main rigid portion and articulated portion. It is more complicated that “rigid” object in that the relative positions, shapes or angles between the main portion and the articulated portion have essentially infinite variations, in addition to the infinite variations of each individual rigid portions due to orientations, rotations and topological transformations. A new method generalized from linear combination is employed to investigate such problems. It uses very few learning samples, and can describe, understand, and recognize 3D articulated objects while the objects status is being changed in an active vision environment.

Moments are widely used in pattern recognition, image processing, computer vision and multiresolution analysis. To clarify and to guide the use of different types of moments, we present in this paper a study on the different moments and compare their behavior. After an introduction to geometric, Legendre, Hermite and Gaussian–Hermite moments and their calculation, we analyze at first their behavior in spatial domain. Our analysis shows orthogonal moment base functions of different orders having different number of zero-crossings and very different shapes, therefore they can better separate image features based on different modes, which is very interesting for pattern analysis and shape classification. Moreover, Gaussian–Hermite moment base functions are much more smoothed, they are thus less sensitive to noise and avoid the artifacts introduced by window function discontinuity. We then analyze the spectral behavior of moments in frequency domain. Theoretical and numerical analyses show that orthogonal Legendre and Gaussian–Hermite moments of different orders separate different frequency bands more effectively. It is also shown that Gaussian–Hermite moments present an approach to construct orthogonal features from the results of wavelet analysis. The orthogonality equivalence theorem is also presented. Our analysis is confirmed by numerical results, which are then reported.

The images processed by our brain represent our window into the world. For some animals this window is derived from a single eye, for others, including humans, two eyes provide stereo imagery, for others like the black widow spider several eyes are used (8 eyes), and some insects like the common housefly utilize thousands of eyes (ommatidia). Still other animals like the bat and dolphin have eyes for regular vision, but employ acoustic sonar vision for seeing when their regular eyes do not work such as in pitch black caves or turbid water. Of course, other animals have adapted to dark environments by bringing along their own lighting such as the firefly and several creatures from the depths of the ocean floor. Animal vision is truly varied and has developed over millennia in many remarkable ways. We have learned a lot about vision processes by studying these animal systems and can still learn even more.

This paper discusses the research in small target detection in infrared images with heavy clutter background. For most infrared images, ship objects are rather dim in the relative dark sea surface background. The existence of scan line disturbance and noise also increases the difficulty in proper detection. Dim objects must be distinguished from a dark background. On the other hand, the small targets must also be distinguished from clutters. Through analysis of the targets and background, we build characteristic models of small ship objects, noise and sea backgrounds respectively, and indicate their differences in spatial and frequency domains among them. Based on the principles of signal processing, pattern recognition and artificial intelligence, we propose a combined algorithm for detecting sea surface small targets. In this algorithm, components of background and noise are first suppressed by a multilevel filter designed accordingly, meanwhile enhancing the target ones of interest. The pixels of the candidate targets are then discriminated by minimum risk Bayes test. Finally, according to a priori knowledge about the targets such as the ranges of their sizes, the targets of interest can be detected. In particular, the related probability distributions used by statistic decision are obtained by offline learning of typical training samples. Experiments show that the algorithm is excellent for such kinds of target detection and is robust to noise.

The Minimum Description Length (MDL) criterion is used to fit a facet model of a car to an image. The best fit is achieved when the difference image between the car and the background has the greatest compression. MDL overcomes the overfitting and parameter precision problems which hamper the more usual maximum likelihood method of model fitting. Some preliminary results are shown.

In this paper an integrated vision system for autonomous land vehicle is described. The vision system includes 2D and 3D vision modules and information fusion module. The task of 2D vision is to provide the physical and geometry information of road, and the task of 3D vision system is to detect the obstacles in the surrounding. Fusion module combines 2D and 3D information to generate a feasible region provided for vehicle navigation.

This paper proposes a general formalism for representation, inference and learning with general hybrid Bayesian networks in which continuous and discrete variables may appear anywhere in a directed acyclic graph. The formalism fuzzifies a hybrid Bayesian network into two alternative forms: the first form replaces each continuous variable in the given directed acyclic graph (DAG) by a partner discrete variable and adds a directed link from the partner discrete variable to the continuous one. The mapping between two variables is not crisp quantization but is approximated (fuzzified) by a conditional Gaussian (CG) distribution. The CG model is equivalent to a fuzzy set but no fuzzy logic formalism is employed. The conditional distribution of a discrete variable given its discrete parents is still assumed to be multinomial as in discrete Bayesian networks. The second form only replaces each continuous variable whose descendants include discrete variables by a partner discrete variable and adds a directed link from that partner discrete variable to the continuous one. The dependence between the partner discrete variable and the original continuous variable is approximated by a CG distribution, but the dependence between a continuous variable and its continuous and discrete parents is approximated by a conditional Gaussian regression (CGR) distribution. Obviously, the second form is a finer approximation, but restricted to CGR models, and requires more complicated inference and learning algorithms. This results in two general approximate representations of a general hybrid Bayesian networks, which are called here the fuzzy Bayesian network (FBN) form-I and form-II. For the two forms of FBN, general exact inference algorithms exists, which are extensions of the junction tree inference algorithm for discrete Bayesian networks. Learning fuzzy Bayesian networks from data is different from learning purely discrete Bayesian networks because not only all the newly converted discrete variables are latent in the data, but also the number of discrete states for each of these variables and the CG or CGR distribution of each continuous variable given its partner discrete parents or both continuous and discrete parents have to be determined.

This paper describes a new system for extracting and classifying bibliography regions from the color image of a book cover. The system consists of three major components: preprocessing, color space segmentation and text region extraction and classification.

Preprocessing extracts the edge lines of the book and geometrically corrects and segments the input image, into the parts of front cover, spine and back cover.

The same as all color image processing researches, the segmentation of color space is an essential and important step here. Instead of RGB color space, HSI color space is used in this system. The color space is segmented into achromatic and chromatic regions first; and both the achromatic and chromatic regions are segmented further to complete the color space segmentation.

Then text region extraction and classification follow. After detecting fundamental features (stroke width and local label width) text regions are determined. By comparing the text regions on front cover with those on spine, all extracted text regions are classified into suitable bibliography categories: author, title, publisher and other information, without applying OCR.

A new scheme of radar target recognition based on parameterized high resolution range profiles (PHRRP) is presented in this paper. A novel criterion called generalized-weighted-normalized correlation (GWNC) is proposed for measuring the similarity between PHRRP's. By properly choosing the parameter of the mainlobe width in GWNC, aspect sensitivity of PHRRP's can be reduced without sacrificing their discriminative power. Performance of the scheme is evaluated using a dataset of three scaled aircraft models. The experimental results show that by using GWNC, only a small number of most dominant scatterers can achieve the same recognition rates as HRRP's, thus leading to a significant data reduction for the recognition system.