Image Segmentation Step Research Articles

Template-based quadrilateral mesh generation from imaging data

This paper describes a novel template-based meshing approach for generating good quality quadrilateral meshes from 2D digital images. This approach builds upon an existing image-based mesh generation technique called Imeshp, which enables us to create a segmented triangle mesh from an image without the need for an image segmentation step. Our approach generates a quadrilateral mesh using an indirect scheme, which converts the segmented triangle mesh created by the initial steps of the Imesh technique into a quadrilateral one. The triangle-to-quadrilateral conversion makes use of template meshes of triangles. To ensure good element quality, the conversion step is followed by a smoothing step, which is based on a new optimization-based procedure. We show several examples of meshes generated by our approach, and present a thorough experimental evaluation of the quality of the meshes given as examples.

영상에서 웨이블렛 기반 로컬 히스토그램 분석을 이용한 에지검출

Edge detection in images is an important step in image segmentation and object recognition as preprocessing for image processing. This paper presents a new edge detection using local histogram analysis based on wavelet transform. In this work, the wavelet transform uses three components (horizontal, vertical and diagonal) to find the magnitude of the gradient vector, instead of the conventional approach in which tw components are used. We compare the magnitude of the gradient vector with the threshold that is obtained from a local histogram analysis to conclude that an edge is present or not. Some experimental results for our edge detector with a Sobel, Canny, Scale Multiplication, and Mallat edge detectors on sample images are given and the performances of these edge detectors are compared in terms of quantitative and qualitative measures. Our detector performs better than the other wavelet-based detectors such as Scale Multiplication and Mallat detectors. Our edge detector also preserves a good performance even if the Sobel and Canny detector are sharply low when the images are highly corrupted.

Partial Occlusion Modulates Contour-Based Shape Encoding in Primate Area V4

Past studies of shape coding in visual cortical area V4 have demonstrated that neurons can accurately represent isolated shapes in terms of their component contour features. However, rich natural scenes contain many partially occluded objects, which have "accidental" contours at the junction between the occluded and occluding objects. These contours do not represent the true shape of the occluded object and are known to be perceptually discounted. To discover whether V4 neurons differentially encode accidental contours, we studied the responses of single neurons in fixating monkeys to complex shapes and contextual stimuli presented either in isolation or adjoining each other to provide a percept of partial occlusion. Responses to preferred contours were suppressed when the adjoining context rendered those contours accidental. The observed suppression was reversed when the partial occlusion percept was compromised by introducing a small gap between the component stimuli. Control experiments demonstrated that these results likely depend on contour geometry at T-junctions and cannot be attributed to mechanisms based solely on local color/luminance contrast, spatial proximity of stimuli, or the spatial frequency content of images. Our findings provide novel insights into how occluded objects, which are fundamental to complex visual scenes, are encoded in area V4. They also raise the possibility that the weakened encoding of accidental contours at the junction between objects could mark the first step of image segmentation along the ventral visual pathway.

SEGMENTASI CITRA SECARA SEMI-OTOMATIS UNTUK VISUALISASI VOLUMETRIK CITRA CT-SCAN PELVIS

Semi-Automatic Image Segmentation for Volumetric Visualization of Pelvis CT Scan-Images. The current development of computerized tomography (CT) has enable us to obtain cross sectional image using multi slicing techniques in an order of few seconds. The obtained images represent several tissue structures on cross section slice being imaged. One challenge to help diagnosis using CT images is extracting an anatomic structure of interest using a method of image segmentation and volumetric visualization with the assistance of computers. In case of volumetric visualization of pelvis bones extracted from multi-slice CT images, whole images which are containing part of pelvis bone structures must be segmented. In this research, an image segmentation technique based on active contour is implemented for semi-automatic multi slice image segmentation. Image segmentation steps are initialized with a define model of 2D curve on the first slice image manually. Next, its model curve is deformed to reach the final result of 2D curve that fits to boundary edges of pelvis bone image. The final result of 2D curve on previous slice image was used as an initialization model of 2D curve on the next slice images. This process will continue until the final slice image. This segmentation method was compared with the segmentation method based on threshold from homogenous intensity distribution and manual segmentation method. Quantitative analysis from the results of segmentation on each slice and qualitative analysis on the representation of volumetric visualization are performed in this research.

Geometric modelling of heterogeneous and complex foods

A procedure to obtain realistic geometric models of foods having different inner tissues or sub-regions is developed. The proposed methodology consists in colour segmentation of food images using a distance criterion, obtaining a reduced set of pixels representing univocally all boundaries of food sub-regions, and finally construction of the geometric model through linear interpolation. The procedure was applied to samples of different nature and complexity. The geometric models were assessed in two different ways, i.e. evaluating the performance of the image segmentation step and simulating a chilling process. The former provided an objective assessment while the later verified the usefulness of the geometric models. An optimized scenario was found between the approximation degree of the food boundaries and the computational resources involved in process simulation. Furthermore, the presented procedure can be used to perform food quality evaluation.

In situ particle size estimation for crystallization processes by multivariate image analysis

Crystal size estimation from in situ images has received attention recently as a means to estimate product properties in real-time. In this paper, an automated image analysis strategy that combines classical image analysis techniques with multivariate statistics has been developed for online analysis of in situ images from crystallization process. The strategy introduces a novel image segmentation step based on information extracted from multivariate statistical models. Experimental results for batch cooling crystallization of monosodium glutamate show that the strategy effectively extracts crystal size and shape information from in situ images. The robustness and efficiency of the method has been established by comparing its performance with those obtained by manual analysis of the images. The method yields reasonably good estimates of particle length and is also fast enough to provide online measurements for the purpose of online optimization and control of a typical crystallization process.

Improved atomic meshes

AbstractThis work describes several improvements over the technique known as atomic meshes introduced by Hale (Tenth International Meshing Roundtable, Newport Beach, CA, 2001), which aims at producing numerical meshes from image data ‘directly’, i.e. bypassing any image segmentation step. We propose two modifications to Hale's original idea: (1) a procedure whereby atoms are initially projected onto perceived features, and (2) the introduction of a Laplacian component to the formulation of the energy field governing atom movement. These aim at improving the adaptation of the mesh to the image while producing better‐shaped triangles. True to Hale's idea, no image segmentation is ever tried, but since the atom placement and movement criteria are more sensitive to image borders, experiments indicate that the resulting meshes tend to conform more closely to the image features. Copyright © 2007 John Wiley & Sons, Ltd.

A high-performance VLSI architecture for the histogram peak-climbing data clustering algorithm

Image feature separation is a crucial step for image segmentation in computer vision systems. One efficient and powerful approach is the unsupervised clustering of the resulting data set; however, it is a very computationally intensive task. This paper presents a high-performance architecture for unsupervised data clustering. This architecture is suitable for VLSI implementations. It exploits paradigms of massive connectivity like those inspired by neural networks, and parallelism and functionality integration that can be afforded by emerging nanometer semiconductor technologies. By utilizing a "global-quasi-systolic, local-hyper-connected" architectural approach, the hardware can process real-time DVD-quality video at the highest rate allowed by the MPEG-2 standard. The architecture is a realization of the histogram peak-climbing clustering algorithm, and it is the first special-purpose architecture that has been proposed for this important problem. The architecture has also been prototyped using a Xilinx field programmable gate array (FPGA) development environment. Although this paper discusses a computer vision application, the architecture presented can be utilized in the acceleration of the clustering process of any type of high-dimensionality data.

Image segmentation for automatic particle identification in electron micrographs based on hidden Markov random field models and expectation maximization

Three-dimensional reconstruction of large macromolecules like viruses at resolutions below 10 Å requires a large set of projection images. Several automatic and semi-automatic particle detection algorithms have been developed along the years. Here we present a general technique designed to automatically identify the projection images of particles. The method is based on Markov random field modelling of the projected images and involves a pre-processing of electron micrographs followed by image segmentation and post-processing. The image is modelled as a coupling of two fields—a Markovian and a non-Markovian. The Markovian field represents the segmented image. The micrograph is the non-Markovian field. The image segmentation step involves an estimation of coupling parameters and the maximum á posteriori estimate of the realization of the Markovian field i.e, segmented image. Unlike most current methods, no bootstrapping with an initial selection of particles is required.

Hyperspectral edge filtering for measuring homogeneity of surface cover types

Edge operators are widely used on grey-level images as a first step in image segmentation or image interpretation. Methods that are currently used for edge filtering are complex and only valid under certain conditions, or useful for three-band colour images only. The problem remains on how to apply edge filtering on multispectral or even hyperspectral images. This paper presents a method that can be used for multispectral and hyperspectral edge filtering. The method uses filtering techniques in combination with three distance measures, namely the Euclidean distance, the spectrum intensity difference, or the spectral angle. The same hyperspectral edge filters serve as homogeneity measures. By simply ‘plugging in’ another distance measure in the same filter, a separation between intensity and spectral content is achieved, which makes the hyperspectral edge filters the natural extension of colour edge filters. The hyperspectral edge filters use information from all spectral bands to arrive at a (dis)similarity map in one step. The edge filters were tested on HyMap hyperspectral imagery of the island of Schiermonnikoog. The test area includes agriculture and a saltmarsh. Experimental results are presented. The spectral angle (dis)similarity of neighbouring pixels can be directly compared with spectral angles already obtained by using the Spectral Angle Mapper for classification. We show that these hyperspectral edge filters assist image interpretation, even in heterogeneous ecosystems such as saltmarshes.

A pseudo top-hat mathematical morphological approach to edge detection in dark regions

Edge detection is an important pre-processing step in image segmentation. Conventionally, edges are detected according to gradient property, then processed by the thresholding technique. By such an approach, fine edge details in dark region of the image are eliminated. It is annoying sometimes as they are as useful as those in bright region, although this is caused by unevenly distributed lighting in many machine vision applications. In this paper, a novel mathematical morphological edge detection algorithm, based on pseudo top-hat transformation which is derived from top-hat transformation, is proposed to preserve these edge details as well as prominent ones. The algorithm is also presented in detail. Comprehensive experimental results show that the proposed algorithm is efficient for edge details extraction in place of shading while preserving distinguish features.

Fractal-based classification of natural textures

Texture classification is an important first step in image segmentation and image recognition. The classification algorithm must be able to overcome distortions, such as scale, aspect and rotation changes in the input texture. In this paper, a new fractal model for texture classification is presented. The model is based on fractional Brownian motion (FBM). It is also shown that this model is invariant to changes in incident light; empirical results are also given. The isotropic nature of Brownian motion is particularly useful for outdoor applications, where the viewing direction may change. Classification results of this model are presented; comparisons with other texture measurement models indicate that the incremental FBM (IFBM) model has better performance for the samples tested.

A Mathematical Morphology Approach to the Star/Galaxy Characterization

The automated star/galaxy discrimination problem is a broadly studied issue of astronomical imaging. The success of a discrimination task depends on the features selected to characterize both classes of interest. In this work we propose an original approach to the characterization of these astronomical objects through the use of Mathematical Morphology, based on gray-level shape-size information. Our method consists of image preprocessing, segmentation and feature extraction steps, all of which employ Mathematical Morphology tools that were implemented using the MMach toolbox for the Khoros system. We briefly present a comparison between our segmentation results, based on the watershed method, and those of the SExtractor software package. The shape-size features are extracted through the use of the gray-level morphological pattern spectrum, which yields attractive analysis attributes that promise to be very suitable for future work in neural-network-based automatic classification

Fuzzy random fields and unsupervised image segmentation

Statistical unsupervised image segmentation using fuzzy random fields is treated. A fuzzy model containing a hard component, which describes pure pixels, and a fuzzy component which describes mixed pixels, is introduced. A procedure for simulating, a fuzzy field based on a Gibbs sampler step followed by a second step involving white or correlated Gaussian noises is given. Then the different steps of unsupervised image segmentation are studied. Four different blind segmentation methods are performed: the conditional expectation, two variants of the maximum likelihood, and the least squares approach. The parameters required are estimated by the stochastic estimation maximization (SEM) algorithm, a stochastic variant of the expectation maximization (EM) algorithm. These fuzzy segmentation methods are compared with a classical hard segmentation method, without taking the fuzzy class into account. The study shows that the fuzzy SEM algorithm provides reliables estimators. Furthermore, fuzzy segmentation always improves upon the hard segmentation results. >