This paper presents a methodology for evaluation of low-level image analysis methods, using binarization (two-level thresholding) as an example. Binarization of scanned gray scale images is the first step in most document image analysis systems. Selection of an appropriate binarization method for an input image domain is a difficult problem. Typically, a human expert evaluates the binarized images according to his/her visual criteria. However, to conduct an objective evaluation, one needs to investigate how well the subsequent image analysis steps will perform on the binarized image. We call this approach goal-directed evaluation, and it can be used to evaluate other low-level image processing methods as well. Our evaluation of binarization methods is in the context of digit recognition, so we define the performance of the character recognition module as the objective measure. Eleven different locally adaptive binarization methods were evaluated, and Niblack's method gave the best perf...

Abstract—A hybrid multidimensional image segmentation algorithm is proposed, which combines edge and region-based techniques through the morphological algorithm of watersheds. An edge-preserving statistical noise reduction approach is used as a preprocessing stage in order to compute an accurate estimate of the image gradient. Then, an initial partitioning of the image into primitive regions is produced by applying the watershed transform on the image gradient magnitude. This initial segmentation is the input to a computationally efficient hierarchical (bottomup) region merging process that produces the final segmentation. The latter process uses the region adjacency graph (RAG) representation of the image regions. At each step, the most similar pair of regions is determined (minimum cost RAG edge), the regions are merged and the RAG is updated. Traditionally, the above is implemented by storing all RAG edges in a priority queue. We propose a significantly faster algorithm, which additionally maintains the so-called nearest neighbor graph, due to which the priority queue size and processing time are drastically reduced. The final segmentation provides, due to the RAG, one-pixel wide, closed, and accurately localized contours/surfaces. Experimental results obtained with two-dimensional/three-dimensional (2-D/3-D) magnetic resonance images are presented. Index Terms — Image segmentation, nearest neighbor region merging, noise reduction, watershed transform. I.

In this paper a multiple frame based technique to enhance text in digital video is presented. After extracting a reference text block, we use an image matching technique to find the corresponding text blocks in consecutive frames. We register these text blocks to subpixel levels by using image interpolation techniques to improve both correspondence and text resolution. The registered text blocks are averaged to obtain a new text block with a clean background and a higher resolution. Experiments conducted on several video sequences show that our enhancement scheme can improve the accuracy of commercial off-the-shelf OCR considerably. 1 Introduction Text extraction and recognition in digital video has drawn more and more attention in recent years as a viable way to provide supplemental index information for a variety of tasks. Compared with the text recognition problem in typical document images, text recognition in video presents several new challenges. First, text in video usually ha...

We consider the problem of segmenting a digitized 2D or 3D image consisting of two univariate populations. Assume a-priori knowledge allows incomplete assignment of voxels in the image, in the sense that a fraction of the voxels can be identified as belonging to population \Pi 0 , a second fraction to \Pi 1 , and the remaining fraction have no a-priori identification. Based upon estimates of the short length scale spatial covariance of the image, we develop a method utilizing indicator kriging to complete the image segmentation. Keywords--- Image segmentation, spatial thresholding, indicator kriging I. Introduction We are interested in the problem of image segmentation by thresholding for the two population univariate case when an image consists of an object (population \Pi 1 ), of possibly complicated shape, and a background (population \Pi 0 ). Our work is motivated by three dimensional X-ray tomographic (CAT) or laser scanning confocal microscopic (LSCM) images of biphase material...

. We present spatial distributions for pore path length and coordination number, pore throat size and nodal pore volume obtained for a 1:5 3 mm 3 volume of 12.1% porosity Fontainebleau sandstone. The sandstone was imaged using Synchrotron X-Ray computed microtomography at 6 micron resolution. The spatial distributions were computed based upon three dimensional medial axis analysis of the void space in the image. We also present vesicle size and vesicle-vesicle contact surface area distributions for a 1.36 mm length of a 6.36 mm diameter core of basalt from a vesiculated lava flow imaged at 20 micron resolution. 1 Introduction The leading X-ray computed microtomographic(CAT) imaging facilities can now provide 1024 3 voxel images of rock and other porous media samples at a voxel resolution of just under 6 microns (Spanne et al., 1994). Scanning laser confocal microscopy (LSCM) measurements of rock can provide images at sub-micron resolution, though with limitations involving depth...

One difficulty with using text from digital video for indexing and retrieval is that video images are often in low resolution and poor quality, and as a result, the text can not be recognized adequately by most commercial OCR software. Text image enhancement is necessary to achieve reasonable OCR accuracy. Our enhancement consists of two main procedures, resolution enhancement based on Shannon interpolation and text separation from complex image background. Experiments show our enhancement approach improves OCR accuracy considerably. Keywords: Text enhancement, Adaptive thresholding, OCR, Shannon Interpolation 1. INTRODUCTION The increasing availability of online digital imagery and video has rekindled interest in the problems of how to index multimedia information sources automatically and how to browse and manipulate them efficiently. Although content based recognition has not progressed to the point where it is useful for indexing heterogeneous collections, text in digital videos...

The 3DMA code is designed to provide statistical analysis of the geometrical distribution of the phases in a two or three dimensional image of a bi-phase material. The code has been applied to the analysis of void and grain phases of rock, cellulose fiber networks, fish movement, and the structure of neurons. For literature studies based upon the 3DMA code see references [4, 5, 6]. This guide provides general installation and execution instructions for the 3DMA code, version Dec. 1999. In addition to general algorithms employed in all four applications, algorithms to support specialized computation for each of these applications are also contained in the Dec. 99 version. This manual describes only the general algorithms and those designed for applications analogous to that for rock microstructure. Supported by the Geosciences Program of the U.S. Department of Energy, grant DEFG02 -92ER14261; the Applied Mathematics Subprogram of the U.S. Department of Energy, grant DE-FG02-...

of the Dissertation Random Field Simulation and an Application of Kriging to Image Thresholding by Wonho Oh Doctor of Philosophy in Applied Mathematics and Statistics State University of New York at Stony Brook 1998 This thesis contains two parts. In Part I we develop a parallel algorithm to generate realizations in a rectangular R 2 or R 3 domain of a stochastic, isotropic, scalar field which is conditioned pointwise to a set of measurements using a kriging procedure. The field is characterized by heterogeneity variation described either by a two point covariance function C(r) or semivariogram fl(r) for pairs of points separated by distance r. We describe the implementation of the algorithms and present numerical examples with discussion. In Part II we apply the theory of kriging to the problem of image segmentation. Consider a digitized (2D or 3D) image consisting of two univariate populations. Asiii sume a-priori knowledge allows incomplete assignment of voxels in the image, in...

The three-dimensional geometry and connectivity of pore space controls the hydraulic transport behavior of crustal rocks. We report on direct measurement of flow-relevant geometrical properties of the void space in a suite of 4 samples of Fontainebleau sandstone ranging from 7.5% to 22% porosity. The measurements are obtained from computer analysis of three dimensional, synchrotron X-ray computed microtomographic images. We present measured distributions of coordination number, channel length, throat size and pore volume, and of correlations between throat-size/pore-volume and nearest neighbor pore-volume/pore volume determined for these samples. We also present quantitative characterization of the distributions measured. The effects of finite sample volume are investigated. The accuracy of the numerical algorithms employed is investigated using a simulated image of hexagonal closed packed spheres.

Document binarization is an active research area for many years. There are many difficulties associated with satisfactory binarization of document images and especially in cases of degraded historical documents. In this paper, we try to answer the question “how well an existing binarization algorithm can binarize a degraded document image? ” We propose a new technique for the validation of document binarization algorithms. Our method is simple in its implementation and can be performed on any binarization algorithm since it doesn’t require anything more than the binarization stage. Then we apply the proposed technique to 30 existing binarization algorithms. Experimental results and conclusions are presented. 1.