Showing papers on "Segmentation-based object categorization published in 1986"

Graph theory for image analysis: an approach based on the shortest spanning tree

Abstract: The paper describes methods of image segmentation and edge detection based on graph-theoretic representations of images. The image is mapped onto a weighted graph and a spanning tree of this graph is used to describe regions or edges in the image. Edge detection is shown to be a dual problem to segmentation. A number of methods are developed, each providing a different segmentation or edge detection technique. The simplest of these uses the shortest spanning tree (SST), a notion that forms the basis of the other improved methods. These further methods make use of global pictorial information, removing many of the problems of the SST segmentation in its simple form and of other pixel linking algorithms. An important feature in all of the proposed methods is that regions may be described in a hierarchical way.

Segmentation of a Thematic Mapper Image Using the Fuzzy c-Means Clusterng Algorthm

Abstract: In this paper, a segmentation procedure that utilizes a clustering algorithm based upon fuzzy set theory is developed. The procedure operates in a nonparametric unsupervised mode. The feasibility of the methodology is demonstrated by segmenting a six-band Landsat-4 digital image with 324 scan lines and 392 pixels per scan line. For this image, 100-percent ground cover information is available for estimating the quality of segmentation. About 80 percent of the imaged area contains corn and soybean fields near the peak of their growing season. The remaining 20 percent of the image contains 12 different types of ground cover classes that appear in regions of diffferent sizes and shapes. The segmentation method uses the fuzzy c-means algorithm in two stages. The large number of clusters resulting from this segmentation process are then merged by use of a similarity measure on the cluster centers. Results are presented to show that this two-stage process leads to separation of corn and soybean, and of several minor classes that would otherwise be overwhelmed in any practical one-stage clustering.

Statistical model-based algorithms for image analysis

Abstract: In this paper, two-dimensional stochastic linear models are used in developing algorithms for image analysis such as classification, segmentation, and object detection in images characterized by textured backgrounds. These models generate two-dimensional random processes as outputs to which statistical inference procedures can naturally be applied. A common thread throughout our algorithms is the interpretation of the inference procedures in terms of linear prediction residuals. This interpretation leads to statistical tests more insightful than the original tests and makes the procedures computationally tractable. This paper also examines a computational structure tailored to one of the algorithms. In particular, we describe a processor based on systolic arrays that realizes the object detection algorithm developed in the paper.

Models of Segmentation

Abstract: The formation of segmented structures is a very important step during development of higher organisms. With the formation of somites in vertebrates or the segments in insects the primary anteroposterior pattern of the organisms is laid down. Segmentation is the result of the superposition of two pattern formation processes. One generates a periodic pattern, i.e. a repetition of homologous structures. It consists in vertebrates of somites and somitic clefts and in insects of segments and segment borders. Superimposed on this periodic pattern is a sequential pattern which makes the repetitive subunits different from each other. In recent years, we have proposed molecularly feasible models which are able to generate periodic and sequential structures precisely superimposed on each other (Meinhardt, 1982a,b). For insect development more detailed experimental and genetic data are available. For that reason the model of insect segmentation is more advanced. At the beginning of this paper a short overview of the model proposed for insect segmentation will be provided. I will show that, with some adaptations, the basic mechanism is also able to account for somite formation.

Recognition system for run-on handwritten characters

Abstract: A method of processing a word with the segmentation and recognition steps combined into an overall scheme. This is accomplished by a three step procedure. First, potential or trail segmentation points are derived. This is done in a manner so as to ensure that essentially all true segmentation points are present but also includes extra or not true segmentation points. Second, all combinations of the segments that could reasonably be a character are sent to a character recognizer to obtain ranked choices and corresponding scores. Finally, the recognition results are sorted and combined so that the character sequences having the best cummulative scores are obtained as the best word choices. For a particular word choice there is a corresponding character segmentation, simply the segment combinations that resulted in the chosen characters. With this recognition scheme the initial character segmentation is not final and need not be highly accurate, but is subject to a lesser constraint of containing the true segmentation points.

Image Structure Representation and Processing: A Discussion of Some Segmentation Methods in Cytology

Abstract: Image processing methods (segmentation) are presented in connection with a modeling of image structure. An image is represented as a set of primitives, characterized by their type, abstraction level, and a list of attributes. Entities (regions for example) are then described as a subset of primitives obeying particular rules. Image segmentation methods are discussed, according to the associated image modeling level. Their potential efficacity is compared, when applied to cytologic image analysis.

A Pyramid-Based Approach to Segmentation Applied to Region Matching

Abstract: In this paper, we attempt to place segmentation schemes utilizing the pyramid architecture on a firm footing. We show that there are some images which cannot be segmented in principle. An efficient segmentation scheme is also developed using pyramid relinking. This scheme will normally have a time complexity which is a sublinear function of the image diameter, which compares favorably to other schemes. The efficacy of our approach to segmentation using pyramid schemes is demonstrated in the context of region matching. The global features we use are compared to those used in previous approaches and this comparison will indicate that our approach is more robust than the standard moment-based techniques.

A robust method for picture segmentation based on split-and-merge procedure

Abstract: The aim of picture segmentation is the extraction of pertinent and stable areas. Pertinence is the agreement of the detected areas with a physical or semantical property of the object; stability is the robustness of the detection to slight transformations such as geometric or photometric distortions. In aerial picture segmentation, the pertinence of an area is often reduced to radiometric homogeneity and spatial connectivity. Unfortunately stability is seldom checked and the deduced segmentation is very sensitive to many parameters introduced by the programmer and thus it is not very reliable. We propose a solution to the stability problem. It will be presented in a theoretical way and then an example of an application is proposed. This method makes use of the well-known split-and-merge algorithm and we will first recall its principle and its main properties.

An iterative region-growing process for cell image segmentation based on local color similarity and global shape criteria.

Abstract: An image segmentation process was derived from an image model that assumed that cell images represent objects having characteristic relationships, limited shape properties and definite local color features. These assumptions allowed the design of a region-growing process in which the color features were used to iteratively aggregate image points in alternation with a test of the convexity of the aggregate obtained. The combination of both local and global criteria allowed the self-adaptation of the algorithm to segmentation difficulties and led to a self-assessment of the adequacy of the final segmentation result. The quality of the segmentation was evaluated by visual control of the match between cell images and the corresponding segmentation masks proposed by the algorithm. A comparison between this region-growing process and the conventional gray-level thresholding is illustrated. A field test involving 700 bone marrow cells, randomly selected from May-Grunwald-Giemsa-stained smears, allowed the evaluation of the efficiency, effectiveness and confidence of the algorithm: 96% of the cells were evaluated as correctly segmented by the algorithm's self-assessment of adequacy, with a 98% confidence. The principles of the other major segmentation algorithms are also reviewed.

Some pyramid techniques for image segmentation

Abstract: This paper describes a collection of multiresolution, or “pyramid”, techniques for rapidly extracting global structures (features, regions, patterns) from an image. If implemented in parallel on suitable cellular pyramid hardware, these techniques require processing times on the order of the logarithm of the image diameter.

A unified method for segmentation and edge detection using graph theory

Abstract: Methods of image segmentation and edge detection based on graph theoretic representations of images are described. The image is mapped onto a weighted graph and, from this graph, spanning trees are used to describe regions and edges in the image. Edge detection is shown to be a dual problem to segmentation. Two methods are developed for both segmentation and edge detection, providing four related techniques. The simpler method uses the Shortest Spanning Tree (SST) to partition the graph and to form a segmentation or edge detection. The second method applies the first method recursively to incorporate global pictorial information into the graph, removing many problems of the simpler method and of other pixel-linking algorithms. An important property of the segmentation and edge detection methods is that image features may be described in a hierarchical way.

Low Level Segmentation of Noisy Imagery

Abstract: : A possible approach to image segmentation is first to perform a low- level segmentation. This then allows an original image to be described in terms of a set of simple regions or primitives. Objects in the image may be subsequently recognized by matching these primitives to patterns of primitives in a data base. It is found that current techniques for low-level image segmentation fail when applied to high noise images. An algorithm is presented which overcomes the problems associated with high noise and succeeds in generating low-level segmentations of noisy imagery. The algorithm is shown also to work on low noise data.

Quad-Tree based Two Source Image Coding

Abstract: In this paper we describe an adaptation of our two source RBC algorithm[1] where we introduce a variable block size. The blocks are chosen to meet some uniformity criterion and are generated via a quad-tree segmentation of the image which allows the segmentation overhead to be coded very efficiently. The prediction component is improved over standard RBC prediction with a fixed block size and some images may be coded using the prediction alone. However it is more economical to also code the residual component for more detailed images.

Image segmentation: A comparative study

Abstract: Machine extraction of meaningful features from the digitized representation of an image (picture, scene etc.) is of great interest to investigators working in such diverse fields as robotic vision, scene analysis, pattern recognition, and automatic part identification in manufacturing processes. The authors describe in detail their algorithms for implementing different segmentation strategies. These are a label propagation segmentation scheme (using the region growing algorithm) and a linked pyramid segmentation scheme. The two techniques are analyzed and compared with respect to their ability to satisfactorily segment a wide class of images (scenes, radiographs, machine parts etc.); computational overheads; memory overheads; and sensitivity to additive noise (Gaussian). In addition to the critical analysis and evaluation of the two techniques, the authors introduce the following enhancements: new predicates (similarity criteria) that are applicable to a broad class of images; incorporation of impulse noise suppression; hierarchical two-level processing to refine segmentation by label propagation; and use of a weighting function to improve the segmentation process.

A Parallel Algorithm for Stochastic Image Segmentation

Abstract: A parallel algorithm for syntactic image segmentation is introduced. Stochastic tree grammar is used as a context-generating model. It is shown that when this context-generating process is in the equilibrium state, a matched filter can be designed and applied in parallel to the image. This process can be used for image segmentation in a syntactic pattern recognition system to enhance the performance of the succeeding recognition process.

Second Generation Hybrid Image-Coding Techniques

Abstract: Recently, ways to obtain a new generation of image-coding techniques have been proposed. The incorpordtion of the human visual system (IIVS) models and tools of the image analysis, such as segmentation, are two defining features of these techniques. In this paper, an application of the new approach to the classical linear predictive coding (LPC) of images and an HVS based segmentation technique for the second genera-tion coders will be discussed. In the case of LPC, the error image is encoded using an image decomposition approach and binary image coding. This improves the compression ratio keeping the quality nearly the same. The new segmentation technique can be used in single frame image coding applications to obtain acceptable images at extremely high compression rates.

Range image segmentation for object recognition

Abstract: Segmentation is a fundamental first step in recognition of objects in range images, providing region and boundary delineation for image analysis. The simplest boundaries of interest in a range image are step edges, indicated by discontinuities in range. These points are detected by finding directed local maxima in the gradient magnitude, normalized over a small neighborhood. More complex are roof edge boundaries, indicated by a sharp bend in the surface. Sharp bends in the surface along roofs, creases, and at corners are well represented by root-mean-square (rms) curvature, a scalar measure of the surface curvature. Roof edge points are determined by finding a local peaks in the estimated rms curvature. The effect of noise in the image on roof boundary estimation accuracy is lessened by adaptively filtering the range image prior to surface fitting and curvature calculation. A new technique is presented for closing gaps in the boundary map by dilating the boundaries in a reversible manner during the segmentation process. A chamfer map is generated indicating the distance of each pixel from an edge. The image is segmented using an externally specified radius and the chamfer map to close gaps of a small, but arbitrary size. The dilated boundaries are relabeled after segmentation to maximize region area, regaining information lost to the dilated edges for later processes.

Identification of Human Faces Using Data-Driven Segmentation, Rule-based

Abstract: WE PRESENT A SYSTEM FOR IDENTIFYING HUMAN FACES IN GRAY-SCALE TELEVISION IMAGERY. THE SYSTEM USES A THREE-STAGE APPROACH TO IMAGE INTERPRETATION. THE FIRST STAGE IS DATA-DRIVEN IMAGE SEGMENTATION; THE SECOND IS RULE-BASED HYPOTHESIS VERIFICATION USING AN ITERATIVE RELAXATION PROCESS.

Partitioning of texture image using two-dimensional linear prediction model

Abstract: This paper presents a study on texture segmentation based on a two-dimensional linear prediction method. It begins with brief summary of the principle that the division of an image which gives the minimum of the total sum of final prediction errors over the whole image corresponds to the true segmentation. Then, the practical computation procedure is given. Finally, experiments of the application of the principle are demonstrated. The results indicate that arbitrarily-shaped texture regions can be well identified.

Image Segmentation By Cluster Analysis Of High Resolution Textured SPOT Images

Abstract: Textural analysis is now a commonly used technique in digital image processing. In this paper, we present an application of textural analysis to high resolution SPOT satellite images. The purpose of the methodology is to improve classification results, i.e. image segmentation in remote sensing. Remote sensing techniques, based on high resolution satellite data offer good perspectives for the cartography of littoral environment. Textural information contained in the pan-chromatic channel of ten meters resolution is introduced in order to separate different types of structures. The technique we used is based on statistical pattern recognition models and operates in two steps. A first step, features extraction, is derived by using a stepwise algorithm. Segmentation is then performed by cluster analysis using these extracted. features. The texture features are computed over the immediate neighborhood of the pixel using two methods : the cooccurence matrices method and the grey level difference statistics method. Image segmentation based only on texture features is then performed by pixel classification and finally discussed. In a future paper, we intend to compare the results with aerial data in view of the management of the littoral resources.

A New Image Segmentation And Texture Analysis Algorithm

Abstract: Traditionally, image segmentation algorithms work by either making point amplitude measurements or by scanning a small computational window over the scene to discover local texture statistics. Where these measurements significantly change, the boundary of an object is said to exist. In this new image segmentation algorithm the entire scene is first transformed with a global transformation such as with a Fourier or Hadamard transform. The consequence of this transformation is that coherently linked structures within the scene, such as texture fields, condense into one, or a few, distinctive peaks. These peaks may then be selectively extracted (or rejected) by a variety of supplementary algorithms. The result is a modified coherent spectrum of the original scene. Through inverse transformation of this modified spectrum back to the image domain, the coherently linked structures are extracted. With this technique structures of related texture may be selectively, and globally, extracted even if they are not contiguous in the original image -and even in the presence of very substantial noise.

Statistical Models and Methods for Cluster Analysis and Image Segmentation.

Abstract: : Clustering of individuals, segmentation of time series and segmentation of numerical images can all be considered as labeling problems, for each can be described in terms of pairs (x sub t, g sub t), t = 1,2,...,n, where x sub t is the observation at instance t and g sub t is the unobservable label of instance t. The labels are to be estimated, along with any unspecified distributional parameters. In cluster analysis the values of t are the individuals (cases) observed and the x's are independent. In time series the values of t are time instants and there is temporal correlation. In numerical image segmentation the values of t denote picture elements (pixels) and spatial correlation between neighboring pixels can be utilized. The idea in segmentation is that signals and time series often are not homogeneous but rather are generated by mechanisms or processes with various phases. Similarly, images are not homogeneous but contain various objects. Segmentation is a process of attempting to recover automatically the phases or objects. Keywords: Statistical pattern recognition; Classification; Optimization by relaxation method. (Author)

A Study Of Texture Segmentation

Abstract: In recent years, significant progress has been made in image segmentation and classification, but still no global theory of image segmentation exists. The wide variety of segmentation techniques used are basically ad hoc and are very dependent on the way the desired features are presented. The main local features used in various segmentation algorithms are image brightness, color and texture. One of the most important features for image segmentation by the human observer is texture, yet it has been difficult to measure and characterize. Actually, texture segmentation is at a very early stage of development at this time. This paper describes an experimental investigation into digital image segmentation using texture features. Texture features are extracted from a few widely different examples of image data. Several different feature sets are used, and the resulting files are input into an unsupervised clustering algorithm. Several variations on the clustering algorithm are explored: some partition the image into segments by using similarities only in the space of features, and others include spatial information such as the location of individual pixels. The various experimental results are also compared, and a new direction for investigation is described.

Automatic description of microstructures: III. Phase particle classification using shape descriptors

Abstract: SUMMARY Present automatic image analysis techniques quantify a materials morphology in terms of the standard stereological parameters such as constituent size and distribution. While useful this information lacks the ‘depth’ available from qualitative examination. In this first of three papers it is suggested that automatic methods can be developed to capture and describe image structure in the descriptive terminology of a metallographer. The goal of such work is to permit quantification of formerly descriptive parameters. A preliminary stage in image analysis is the segmentation of an image into homogeneous regions in preparation for further processing. It is demonstrated that metallographic images (in this case binary eutectic alloys of aluminium) can be segmented using measures of the spatial organization, or texture, present in an image. A weighted average of the Hadamard transform of an image is shown to give a local measure of the texture suitable for segmentation of the images into homogeneous regions of microstructure.