Abstract: Several model-based algorithms for threshold selection are presented, concentrating on the two-population univariate case in which an image contains an object and background. It is shown how the main ideas behind two important nonspatial thresholding algorithms follow from classical discriminant analysis. Novel thresholding algorithms that make use of available local/spatial information are then given. It is found that an algorithm using alternating mean thresholding and median filtering provides an acceptable method when the image is relatively highly contaminated, and seems to depend less on initial values than other procedures. The methods are also applicable to multispectral k-population images. >

Abstract: The authors investigate the use of a priori knowledge about a scene to coordinate and control bilevel image segmentation, interpretation, and shape inspection of different objects in the scene. The approach is composed of two main steps. The first step consists of proper segmentation and labeling of individual regions in the image for subsequent ease in interpretation. General as well as scene-specific knowledge is used to improve the segmentation and interpretation processes. Once every region in the image has been identified, the second step proceeds by testing different regions to ensure they meet the design requirements, which are formalized by a set of rules. Morphological techniques are used to extract certain features from the previously processed image for rule verification purposes. As a specific example, results for detecting defects in printed circuit boards are presented. >

Abstract: For pt.I see ibid., vol.9, no.6, p.787 (1987). The problem of uncertainty in image feature description is discussed, and it is shown how finite prolate spheroidal sequences can be used in the construction of feature descriptions that combine spatial and frequency-domain locality in an optimal way. Methods of constructing such optimal feature sets, which are suitable for graphical implementation, are described, and some generalizations of the quadtree concept are presented. These methods are illustrated by examples from image processing applications, including feature extraction and texture description. The problem of image segmentation is discussed, and the importance of scale invariance in overcoming the limitations imposed by uncertainty is demonstrated. A novel texture segmentation algorithm that is based on a combination of the new feature description and multiresolution techniques is described and shown to give accurate segmentations on a range of synthetic and natural textures. >

Abstract: Although the viability of global marketing is disputed, the best opportunities for pursuing basically the same strategy across national borders are in industrial marketing. However, because of the disparities across world markets, segmentation is essential to assessing opportunities for a standardised marketing approach. Segmentation based on economic indicators represents the first step in identifying potential markets. In this study, 96 countries were grouped into six segments. Implications for industrial marketers are presented, along with issues relating to using stages of economic development as a basis for segmentation and using a factor analytic and clustering approach to the segmentation of the global market.

Abstract: A range sensing technique is demonstrated that finds the 3-D shape of diffusely reflecting objects. The technique works sequentially in depth direction and is based on structured illumination and focus sensing. A TV camera and analog electronics are used to find the locations in focus of each step of a focus series in TV real time. The depth resolution is not very high, however, the technique is simple, rapid, and well suited to get an overview of a scene in robot vision.

Abstract: This paper describes recent work on acoustic segmentation and phonetic classification as part of an effort in speech understanding system development. The signal representation is based on an auditory model that incorporates known properties of the human auditory system, including critical‐band filtering, saturation, adaptation, forward masking, and synchrony detection. Acoustic landmarks are determined using a measure of local similarity. These landmarks are embedded in a multilevel structure in which information ranging from coarse to fine is represented in an organized fashion. An analysis of the acoustic structure, using 500 utterances from 100 different talkers, shows that it captures over 96% of the acoustic‐phonetic events of interest with an insertion rate of less than 5%. Phonetic classification is achieved by defining a set of generic property detectors based on the knowledge of acoustic phonetics. The settings of the parameters are obtained by a search procedure, using a large body of training ...

Abstract: Extensions to current models for texture segmentation are presented, in which the underlying detector (filter) mechanisms are allowed to adapt to the incoming signal in terms of their dynamical response range and associativities. This system converges on new texton (B. Julesz, 1981) profiles of minimal dimensionality that are used to classify texture regions by a minimum distance classifier in the texture feature space. The three processes of convolution, cooperativity, and classification are individually analyzed and compared with some observations from human texture discrimination experiments. >

Abstract: The problem of segmenting a range image into homogeneous regions in each of which the range data correspond to a different surface is considered. The segmentation sought is a maximum-likelihood (ML) segmentation. Only planes, cylinders, and spheres are considered as presented in the image. The basic approach to segmentation is to divide the range image into windows, classify each window as a particular surface primitive, and group like windows into surface regions. Mixed windows are detected by testing the hypothesis that a window is homogeneous. Homogeneous windows are classified according to a generalized likelihood ratio test which is computationally simple and incorporates information from adjacent windows. Grouping windows of the same surface types is cast as a weighted ML clustering problem. Finally, mixed windows are segmented using an ML hierarchical segmentation algorithm. A similar approach is taken for segmenting visible-light images of Lambertian objects illuminated by a point source at infinity. >

Abstract: A method is presented for segmenting images into a discrete set of classes by first segmenting at low resolution and then progressing to finer resolutions until individual pixels are classified. This multiple resolution method results in accurate segmentations and requires significantly less computation than some previously known methods. The segmentation algorithm used at each resolution is based on maximum a posteriori estimation of the field of pixel classifications, which is modeled as a Markov random field. The maximization is performed by a deterministic greedy algorithm which iteratively chooses the classification of individual pixels or blocks of pixels. A texture model is also developed which allows the extraction of a texture statistic for each pixel and is well suited for use with the proposed algorithm. Measurements of algorithm performance under varying conditions of region size and signal-to-noise ratio are presented. >

Abstract: In this paper, we present an approach to color image understanding that can be used to segment and analyze surfaces with color variations due to highlights and shading. We begin with a theory -the Dichromatic Reflection Model - that relates the reflected light from dielectric materials, such as plastic, to fundamental physical reflection processes, and describes the color of the reflected light as a linear combination of the color of the light due to surface reflection (highlights) and body reflection (object color). This dichromatic theory is used in an algorithm that separates a color image into two parts: an image of just the highlights, and the original image with the highlights removed. In the past, we have applied this method to hand- segmented images. This paper shows how to perform automatic segmentation by applying this theory in stages to identify the object and highlight colors, The result is a combination of segmentation and reflection analysis that is better than traditional heuristic segmentation methods and provides important physical information about the surface geometry and material properties at the same time. This line of research can lead to physics-based image understanding methods that are both more reliable and more useful than traditional methods.

Abstract: A statistical method of segmentation using a hidden Markov model (HMM) and a Bayesian classifier is described. The main features of this method are the use of feature parameters which are independent of each category in vowels of consonants, and the use of only one HMM which commonly represents all syllable patterns. The segmentation strategy is to find the optimal HMM sequence. The optimal/best sequence is found by using the O(n) DP matching based on Viterbi algorithm. The concatenated number and boundaries of the best HMM sequence are regarded as the segmentation result. The experimental result on Japanese spoken sentences shows that the rate of segmentation is more than 92% for two male speakers, and the rate is improved to 97.5% by using a duration control mechanism based on a discrete probability distribution. >

Abstract: A coarse segmentation algorithm is presented for segmenting textured images which are composed of regions in each of which the data are modeled as one of C Markov random fields (MRFs). The segmentation sought is a maximum-likelihood (ML) segmentation. The image is partitioned into relatively small disjoint square windows. Each window is examined to see whether it is homogeneous or is mixed, and the texture region(s) that comprises the window is (are) decided by a multiple hypothesis test. The formulation of the complex ML segmentation problem in terms of this simpler window-based multiple-hypothesis problem provides huge computational savings, as ML segmentation is only performed at the windows that fall on the boundary between two regions and with the full knowledge of the two populations that are present in the window. Although the problems and solutions presented are for textured image segmentation, they are extendable to problems such as system identification, speech recognition, and data fusion. >

Abstract: The authors present an approach for integration of manually designed parts in CAD databases. They propose a system to derive spline-based descriptions of various component surfaces of objects given only the range image of the object under consideration. The system consists of two modules. The first module, which is centred on a robust segmentation algorithm, generates a segmentation plot of the given range image. The second module then derives spline-based descriptions for each of the different segmented surfaces. >

Abstract: A means of preserving the perceptually significant features of an image is to extract them by segmentation prior to coding. Distortion can be restricted to less important areas, such as fine detail and texture. This is the basis of several two-component and segmented image coding schemes. The technique is extended to video coding by applying segmentation to the frame difference signal. However, independent processing of each frame by this approach leads to inadequate portrayal of motion. An extension of the method, that incorporates guiding of the segmentation according to the history of the sequence, is developed. Results show a considerable improvement, suggesting that there may be further potential for the application of this approach in low rate video coding. >

Abstract: : In this thesis a new approach to the detection and classification of tactical targets using a multifunction laser radar sensor is developed. Targets of interest were tanks, jeeps, trucks, and other vehicles. Doppler images were segmented by developing a new technique which compensates for spurious doppler returns. Relative range images were segmented using an approach based on range gradients. The resultant shapes in the segmented images were then classified using Zernike moment invariants as shape descriptors. Two classification decision rules were implemented: a classical statistical nearest-neighbor approach and a new biologically-based neural network multilayer perceptron architecture. The doppler segmentation algorithm was applied to a set of 180 real world sensor images. An accurate segmentation was obtained for 89 percent of the images. The new doppler segmentation proved to be a robust method, and the moment invariants were effective in discriminating the tactical targets. Tanks were classified correctly 86 percent of the time. The most important result of this research is the demonstration of the use of a new information processing architecture for military applications. The multilayer perceptron outperformed the nearest-neighbor classifier in every test.

Abstract: The technique presented solves the problem of texture segmentation in two steps. In the first, a textured image is divided into small squares (20*20 in this case) and a hierarchical clustering algorithm related to a choice of ultrametric distances is used to obtain an initial segmentation. In the second step, the texture boundaries are improved using a recursive procedure with vertical and horizontal segments of smaller and smaller size, converging towards the correct texture boundaries. >

Abstract: A knowledge-based system for the segmentation of seismic sections is presented. The system can be functionally divided into a texture feature extraction part and a knowledge-based segmentation part. An important characteristic of the proposed approach is the iterative quadtree splitting (IQS) scheme used to control the segmentation process. The final output of the system is a segmentation of the input section into regions (segments) of common signal character. Test runs of the system on a real seismic section from the Gulf of Mexico show that the introduction of domain expert geologic knowledge can significantly improve the overall segmentation. The IQS control scheme provides two functions essential to most knowledge-based image processing and interpretation systems: (1) the coordination of all parallel-operated processes over the entire section for an overall balanced result; and (2) the incorporation of various types of knowledge into the different levels of decision-making in those processes. >

Abstract: Describes an algorithm that performs speaker-independent segmentation and broad classification of continuous speech. The algorithm is implemented as a set of knowledge sources that apply rules to speech parameters to locate segment boundaries and assign broad category labels to the resulting segments. The output of the algorithm is a network of segments with broad category labels. The structure of the algorithm, the manner in which acoustic phonetic knowledge is applied, and the performance of the algorithm on 200 utterances produced by 50 speakers is described. >

Abstract: The segmentation of the velocity vector data into roughly homogeneous blocks which can be represented by a vector is performed in part at least during the prior process of estimation of movement (11). This permits minimisation of the number of components representing the image, as well as the flow of data to control and reconstruct of an image at the decoder. The segmentation uses selection of a limited number of candidate vectors from among the predicted velocity vectors, then subjecting them to a dispersion around their original value before reallocating them in an adaptive 'aggregation/division' process among the image blocks. USE/ADVANTAGE - Optimised compression and faster image generation in sequences of images used in high definition television, e.g. MAC system.

Abstract: Using the Marr-Hildreth operator with a subsequent closing algorithm, it is possible to determine the main constituents of the human head from magnetic resonance image data. With use of the segmentation results in a generalized voxel model one can generate images that are very similar to what is known from real anatomy. It is pointed out that this is facilitated by the tremendous increase of the quality of the original image material. >

Abstract: A rule-based system for image segmentation and understanding is proposed. Called ROF (rule-based object finder), it is the low-level state of an image understanding system for aerial photographs. The general design of ROF uses a novel paradigm for segmentation. Candidate pixels for regions (or borders of regions) are chosen by the raw data module (RDM) on the basis of the geometric shape of their neighborhoods. The RDM has full access to the digital picture and its pixels. Compatible sets of candidate pixels for segment formation are found by search and decision-making techniques by the aggregation module. The segmentation module, the only one that is visible to the user, attempts to construct abstract entities which can serve as global regions or parts of objects in the picture. A specialized module deals with the quantitative side of object features such as size, shade and contrast. This quantification module interfaces to all other modules of ROF. >