Region growing

Page 1. Segmentation of Retinal Blood Vessels Based on the Second Directional Derivative and Region Growing M. Elena Martinez-Pkrez'; Alun D. Hughes2, Alice V. Stanton2, Simon A. Thorn2, Ani1 A. Bharath' and Kim H. Parker' ...

Three dimensional object extraction and recognition (OER) from LIDAR data has been an area of major interest in photogrammetry for quite a long time. However, most of the existing methods for automatic object extraction and recognition from LIDAR data are just based on the range information and employ parametric methods and object’s vagueness behaviour is basically neglected. Thus, these methods do not take into account the extraction and recognition complexities and may fail to reach a satisfied reliability level in complex situations. In this paper a novel approach based on the following strategies is formulated and implemented: (a) for a more comprehensive definition of the objects, information fusion concept is utilized, i.e., object’s descriptive components such as 3D structural and textural (ST) information are automatically extracted from first/last rang and intensity information of LIDAR data and simultaneously fed into the evaluation process, (b) for a more realistic expres...

Computer vision systems attempt to recover useful information about the three-dimensional world from huge image arrays of sensed values. Since direct interpretation of large amounts of raw data by computer is difficult, it is often convenient to partition (segment) image arrays into low-level entities (groups of pixels with similar properties) that can be compared to higher-level entities derived from representations of world knowledge. Solving the segmentation problem requires a mechanism for partitioning the image array into low-level entities based on a model of the underlying image structure. Using a piecewise-smooth surface model for image data that possesses surface coherence properties, we have developed an algorithm that simultaneously segments a large class of images into regions of arbitrary shape and approximates image data with bivariate functions so that it is possible to compute a complete, noiseless image reconstruction based on the extracted functions and regions. Surface curvature sign labeling provides an initial coarse image segmentation, which is refined by an iterative region growing method based on variable-order surface fitting. Experimental results show the algorithm's performance on six range images and three intensity images.

Combining both spatial and intensity information in image, we present an MRI brain image segmentation approach based on multiresolution edge detection, region selection, and intensity threshold methods. The detection of white matter structure in brain is emphasized in this paper. First, a multi-resolution brain image representation and segmentation procedure based on a multi-scale image filtering method is presented. Given the nature of the structural connectivity and intensity homogeneity of brain tissues, region-based methods such as region growing and subtraction to segment the brain tissue structure from the multi-resolution images are utilized. From the segmented structure, the region-of-interest (ROI) image in the structure region is derived, and then a modified segmentation of the ROI based on an automatic threshold method using our threshold selection criterion is presented. Examples on both T1 and T2 weighted MRI brain image segmentation is presented, showing finer brain tissue structures. ᭧

Abstract In this paper an automatic texture based volumetric region growing method for liver segmentation is proposed. 3D seeded region growing is based on texture features with the automatic selection of the seed voxel inside the liver organ and the automatic threshold value computation for the region growing stop condition. Co-occurrence 3D texture features are extracted from CT abdominal volumes and the seeded region growing algorithm is based on statistics in the features space. Each CT volume is composed by 230 slices, ...

A new region growing method for finding the boundaries of blobs is presented. A unique feature of the method is that at each step, at most one pixel exhibits the required properties to join the region. The method uses two novel discontinuity measures, average contrast and peripheral contrast, to control the growing process.

While preparing the composite preform in the powder metallurgy Lab, the various defects due to porosity, open crack and residual stresses are possible. This may lead to poor life and strength of materials. It is difficult to predict the defects in the form of pores physically in the powder metallurgy Lab. To simplify this kind of problem, the Scanning Electron Microscope (SEM) images are generated from the powder composites and are segmented using region growth approach to find the distribution of pores. Normally, the composite preforms are being produced through various processes like mechanical milling, mixing, compaction, sintering and hot extrusion. In this study, Cu–(5–20%) W composite preforms, with a preform density of 94% are prepared. The pore size in term of coverage area, perimeter during different sintering atmospheres are derived. Further, the porosity is reduced during extrusion process. The results of SEM images are compared before and after sintering and extrusion process. This kind of work will aid the manufacturing process of material parts in predicting their strength and life time.

In this paper, we propose an image-based system for Arabic Sign Language (ArSL) recognition. The algorithm starts by detecting the face of the signer using a Gaussian skin color model. The centroid of the detected face is then used as a reference point for tracking the hands' movements. The hands regions are segmented using a region growing algorithm assuming the signer wears a yellow and an orange colored gloves. From the segmented hands regions, an optimal set of features is extracted. To represent the time varying feature patterns, a Hidden Markov Model (HMM) is then used. Before using HMM in testing, the number of states and the number of Gaussian mixtures are optimized. The proposed system was implemented for both signer dependent and signer independent conditions. The experimental results show that an accuracy of more than 95% can be achieved with a large database of 300 signs. The results outperform previous work on ArSL mainly restricted to small vocabulary size.

Sign language recognition is a very challenging research area. In this proposed system, a well trained computer system is used to recognize static hand gestures representing linguistic words. The main aim of the paper is conversion of linguistic sign languages into text and speech form. Recognized sign are also translated into Tamil and Hindi languages. It contains three processes of work. First process is pre-processing, in which the obtained images are processed through the steps like segmentation, resize, and gray conversion. Second process is region-based analysis which exploits both boundary and interior pixels of an object. Solidity, perimeter, convex hull, area, major axis length, minor axis length, eccentricity, orientation are some of the shape descriptors used as features in this process. The features derived are used to train the binary classifier first; secondly the testing images are given for classification. Knn classifiers are used for classification which provides a good result with less computation time for larger datasets. Since, the system handled a binary classifier it performed a one-versus-all kind of classification. PCNN (Pulse Coupled Neural Network) is used for pattern recognition. Third process is the hand gesture recognition.

A new approach for the detection of the road surface and obstacles is presented. The high accuracy of the method allows the detection of traffic isles as distinct class. The 3D data inferred from dense stereo are transformed into a rectangular Digital Elevation Map (DEM). Two classifiers are proposed: density-based and road surface-based. The densitybased obstacle classifier marks DEM cells as road or obstacles, using the density of 3D points as criterion. A quadratic road surface model is initially fitted, by a RANSAC approach, to the region in front of the ego vehicle. A region growing-like process refines this primary solution, driven by the 3D uncertainty model of the stereo sensor. A robust global solution for the road surface is obtained. The road surface is used for discrimination between road, traffic isle and obstacle points. Fusion and error filtering is performed on the results of the two classifiers. The proposed realtime algorithm was evaluated in an urban scenario and can be used in complex applications, from collision avoidance to path planning.

Region growing is a very useful technique for image segmentation. Its efficiency mainly depends on its aggregation criterion. In the present paper, a new algorithm is proposed with a homogeneity criterion based on an adequate tuning between spatial neighbourhood and histogram neighbourhood. It differs from other techniques by reconsidering the pixel (or voxel) assignments on each step by a process which minimizes variance through special dilations. Thus, the region created by an initial seed can be non-connected and possibly does not contain this seed. Examples are given in dental surgery for 2D X-Ray images (and their associated 3D block) and for 3D images acquired by the Morphometre, the new 3D scanner constructed by GEMSE (General Electric Medical Systems). © 1997 Published by Elsevier Science B.V.

Automated extraction of pulmonary anatomy provides a foundation for computerized analysis of computed tomography (CT) scans of the chest. A completely automatic method is presented to segment the lungs, lobes and pulmonary segments from volumetric CT chest scans. The method starts with lung segmentation based on region growing and standard image processing techniques. Next, the pulmonary fissures are extracted by a supervised filter. Subsequently the lung lobes are obtained by voxel classification where the position of voxels in the lung and relative to the fissures are used as features. Finally, each lobe is subdivided in its pulmonary segments by applying another voxel classification that employs features based on the detected fissures and the relative position of voxels in the lobe. The method was evaluated on 100 low-dose CT scans obtained from a lung cancer screening trial and compared to estimates of both interobserver and intraobserver agreement. The method was able to segmen...

c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e 1 1 0 ( 2 0 1 3 ) 150-159 a b s t r a c t Many computer aided diagnosis (CAD) systems help radiologist on difficult task of mass detection in a breast mammogram and, besides, they also provide interpretation about detected mass. One of the most crucial information of a mass is its shape and contour, since it provides valuable information about spread ability of a mass. However, accuracy of shape recognition of a mass highly related with the precision of detected mass contours.

Comic books represent an important cultural heritage in many countries. However, few researches have been done in order to analyse the content of comics such as panels, speech balloons or characters. At first glance, the structure of a comic page may appear easy to determine. In practice, the configuration of the page, the size and the shape of the panels can be different from one page to the next. Moreover, authors often draw extended contents (speech balloon or comic art) that overlap two panels or more. In some situations, the panel extraction can become a real challenge. Speech balloons are other important elements of comics. Full text indexing is only possible if the text can be extracted. However the text is usually embedded among graphic elements. Moreover, unlike newspapers, the text layout in speech balloons can be irregular. Classic text extraction method can fail. We propose, in this paper, a method based on region growing and mathematical morphology to extract automatically the panels of a comic page and a method to detect speech balloons. Our approach is compared with other methods find in the literature. Results are presented and discussed.

A pressure ulcer is a clinical pathology of localized damage to the skin and underlying tissue caused by pressure, shear, or friction. Diagnosis, treatment, and care of pressure ulcers are costly for health services. Accurate wound evaluation is a critical task for optimizing the efficacy of treatment and care. Clinicians usually evaluate each pressure ulcer by visual inspection of the damaged tissues, which is an imprecise manner of assessing the wound state. Current computer vision approaches do not offer a global solution to this particular problem. In this paper, a hybrid approach based on neural networks and Bayesian classifiers is used in the design of a computational system for automatic tissue identification in wound images. A mean shift procedure and a region-growing strategy are implemented for effective region segmentation. Color and texture features are extracted from these segmented regions. A set of multilayer perceptrons is trained with inputs consisting of color and texture patterns, and outputs consisting of categorical tissue classes which are determined by clinical experts. This training procedure is driven by a -fold cross-validation method. Finally, a Bayesian committee machine is formed by training a Bayesian classifier to combine the classifications of the neural networks. Specific heuristics based on the wound topology are designed to significantly improve the results of the classification. We obtain high efficiency rates from a binary cascade approach for tissue identification. Results are compared with other similar machine-learning approaches, including multiclass Bayesian committee machine classifiers and support vector machines. The different techniques analyzed in this paper show high global classification accuracy rates. Our binary cascade approach gives high global performance rates (average sensitivity = 78 7%, specificity = 94 7%, and accuracy = 91 5%) and shows the highest average sensitivity score (=86.3%) when detecting necrotic tissue in the wound.

Recognition of dimensioning text in engineering drawings is an essential part of the drawing understanding process, as this text provides the exact dimensions and tolerances of the object described in the drawing. We consider engineering drawings produced according to either ISO or ANSI drafting standards. Text segmentation and recognition are preceded by orthogonal zigzag vectorization, arc segmentation, and arrowhead pair recognition. Initial textbox extraction is done by a region growing process, performed on text-wire candidates. On the basis of textbox context (neighboring annotation wires) the drafting standard is detected. Raw textboxes are divided into logical textboxes, which are further decomposed into basic textboxes. A neural network based OCR algorithm is applied to each basic textbox. Finally, the OCR recognition results are verified by using contextual information and comparing the results with the measurements made directly on the drawing.

... The reason for the choice of a region growing algorithm as the region-based segmentation technique is that it allows a precise control of the region formation process. The basic technique usually considers only homogeneity conditions such as the distance between the pixel's ...

Reverse engineering is the process of obtaining a geometric CAD model from 3D points acquired by scanning an existing physical model. It is widely used in numerous applications, such as manufacturing, industrial design and jewelry design and ...

A Farsi License Plate Recognition (LPR) System is one kind of automatic inspection of transport systems and is of considerable interest because of its potential applications to areas such as automatic toll collection, traffic law enforcement and security control of restricted areas. This paper proposes an automatic license plate recognition system for Persian license plates. We have different type of Persian license plate with different shape, background, font size and structures. The license plate recognition is done in the several stages. First existence of vehicle detected from desirable windows by image processing algorithms. In the last stages locating of plate in the image is recognize and characters of plate area are extracted by region growing and connectivity algorithms. The end of system are recognized all of characters set of license plate by properly neural network pattern recognition and result stored in database for uses in traffic application. This system worked under variable illumination, variable size of plate and dynamic backgrounds. This system implemented with help of TCT Company and performance of the LPR system has been tested on 400 vehicles images which captured under various sizes of plate and variable illumination conditions. The rate of success recognition is 95%.

Image segmentation is a challenging process in numerous applications. Region growing is one of the segmentation techniques as its basis for the Seeded Region Growing method. A novel real time integrated method is developed in this work to locate the segmented region of interest of an image, based on the Region Growing segmentation method along with the Thresholding supported image segmentation. Through the proposed work, a homogeneity based on pixel intensity is suggested as well as the threshold value can be decided via a variety of schemes such as manual selection, Iterative method, Otsu’s method, local thresholding to obtain the best possible threshold. The experimental results are performed on different images obtained from an Alpert dataset. A comparative study is performed with the human segmented image, threshold based region growing and the proposed integrated method. The results prove that the proposed integrated method outperforms the region growing method in terms of the recall and F-score. While, it has comparable recall values with that gained by the human segmented images. It was noted that as the image under test has a dark background with the brighter object, the proposed method provides the superior recall value compared to the other methods.

Mechanical hardness testing is fundamental in the evaluation of the mechanical properties of metallic materials due to the fact that the hardness values allow one to determine the wear resistance of the material involved, as well as the approximate values of its ductility, flow tension, among a number of other key characteristics. As a result, the main objective of the present work has been the development and analysis of a computational methodology capable of determining the Brinell and Vickers hardness value from hardness indentation images, which is based on image processing and analysis algorithms . In order to validate the methodology which has been developed, comparisons of the results resulting from the consideration of ten indentation image samples obtained through the conventional manual hardness measurement approach and a computational methodology have been carried out.

Historically, anatomical CT and MR images were used to delineate the gross tumour volumes (GTVs) for radiotherapy treatment planning. The capabilities offered by modern radiation therapy units and the widespread availability of combined PET/CT scanners stimulated the development of biological PET imaging-guided radiation therapy treatment planning with the aim to produce highly conformal radiation dose distribution to the tumour. One of the most difficult issues facing PET-based treatment planning is the accurate delineation of target regions from typical blurred and noisy functional images. The major problems encountered are image segmentation and imperfect system response function. Image segmentation is defined as the process of classifying the voxels of an image into a set of distinct classes. The difficulty in PET image segmentation is compounded by the low spatial resolution and high noise characteristics of PET images. Despite the difficulties and known limitations, several image segmentation approaches have been proposed and used in the clinical setting including thresholding, edge detection, region growing, clustering, stochastic models, deformable models, classifiers and several other approaches. A detailed description of the various approaches proposed in the literature is reviewed. Moreover, we also briefly discuss some important considerations and limitations of the widely used techniques to guide practitioners in the field of radiation oncology. The strategies followed for validation and comparative assessment of various PET segmentation approaches are described. Future opportunities and the current challenges facing the adoption of PET-guided delineation of target volumes and its role in basic and clinical research are also addressed.

Breast cancer is regarded as one of the most frequent mortality causes among women. As early detection of breast cancer increases the survival chance, creation of a system to diagnose suspicious masses in mammograms is important. In this paper, two automated methods are presented to diagnose mass types of benign and malignant in mammograms. In the first proposed method, segmentation is done using an automated region growing whose threshold is obtained by a trained artificial neural network (ANN). In the second proposed method, segmentation is performed by a cellular neural network (CNN) whose parameters are determined by a genetic algorithm (GA). Intensity, textural, and shape features are extracted from segmented tumors. GA is used to select appropriate features from the set of extracted features. In the next stage, ANNs are used to classify the mammograms as benign or malignant. To evaluate the performance of the proposed methods different classifiers (such as random forest, naïve Bayes, SVM, and KNN) are used. Results of the proposed techniques performed on MIAS and DDSM databases are promising. The obtained sensitivity, specificity, and accuracy rates are 96.87%, 95.94%, and 96.47%, respectively.

International Regional Science Review serves as an international forum for economists, geographers, planners, and other social scientists to share important research findings and methodological breakthroughs. The journal serves as a catalyst for improving spatial and regional analysis within the social sciences and stimulating communication among the disciplines. IRSR deliberately helps define regional science by publishing key interdisciplinary survey articles that summarize and evaluate previous research and identify fruitful research directions. Focusing on issues of theory, method, and public policy where the spatial or regional dimension is central, IRSR strives to promote useful scholarly research that is securely tied to the real world.

Laser scanning or LiDAR data are increasingly used in forestry applications but also e.g. in urban environments or for building reconstructions. Huge point clouds are usually converted to a grid or are pre-processed in specific software packages. In this paper we present a methodology to extract and delineate single trees from small footprint, high intensity laser scanning point data in

Marginal noise is a common phenomenon in document analysis which results from the scanning of thick documents or skew documents. It usually appears in the front of a large and dark region around the margin of document images. Marginal noise might cover meaningful document objects, such as text, graphics and forms. The overlapping of marginal noise with meaningful objects makes it di cult to perform the task of segmentation and recognition of document objects. This paper proposes a novel approach to remove marginal noise. The proposed approach consists of two steps which are marginal noise detection and marginal noise deletion. Marginal noise detection will reduce an original document image into a smaller image, and then ÿnd marginal noise regions according to the shape length and location of the split blocks. After the detection of marginal noise regions, di erent removal methods are performed. A local thresholding method is proposed for the removal of marginal noise in gray-scale document images, whereas a region growing method is devised for binary document images. Experimenting with a wide variety of test samples reveals the feasibility and e ectiveness of our proposed approach in removing marginal noises.

have developed a recognition system to translate hand gestures into Urdu alphabets. We have formulated a comprehensive classification scheme which is core to this recognition system. The accuracy rate of our approach is 97.4% which is better than the previous approaches.

Accurate quantitative measurements of airway and vascular dimensions are essential for evaluating function in both the normal and in the diseased lung. This report describes a new integrated method for three-dimensional (3D) extraction and analysis of pulmonary tree structures using data from High Resolution Computed Tomography (HRCT). Serially scanned two-dimensional (2D) slices of the lower left Xobe of isolated dog lungs were stacked to create a volume of data. Airway and vascular trees were extracted using a 3D seeded region-growing algorithm based on differences in CT number between wall and lumen. In the region-growing step, voxels in the lumen are tagged with a distance descriptor to identify points along the tree structure equidistant from the seed point. To obtain quantitative data, we reduced each tree to its central axis. From the central axis, branch length was measured as the distance between two successive branch points, branch angle was measured as the angle produced by two daughter branches, and crosssectional area was measured from a plane perpendicular to the central axis point. Data derived from these methods can be used to localize and quantify structural differences both during different physiologic conditions and in pathologic lungs. h'ey Words: HRCT, Lung imaging, Pulmonary tree structures. Bronchial pressure-diameter behavior January_February/l995. Volume 19, Number I Computerized Medical Imaging and Graphic\ January-February/l9YS. Volume 19. Number I

https://www.mdpi.com/1424-8220/21/1/105
This paper introduces a system that can estimate the deformation process of a deformed flat object (folded plane) and generate the input data for a robot with human-like dexterous hands and fingers to reproduce the same deformation of another similar object. The system is based on processing RGB data and depth data with three core techniques: a weighted graph clustering method for non-rigid point matching and clustering; a refined region growing method for plane detection on depth data based on an offset error defined by ourselves; and a novel sliding checking model to check the bending line and adjacent relationship between each pair of planes. Through some evaluation experiments, we show the improvement of the core techniques to conventional studies. By applying our approach to different deformed papers, the performance of the entire system is confirmed to have around 1.59 degrees of average angular error, which is similar to the smallest angular discrimination of human eyes. As a result, for the deformation of the flat object caused by folding, if our system can get at least one feature point cluster on each plane, it can get spatial information of each bending line and each plane with acceptable accuracy. The subject of this paper is a folded plane, but we will develop it into a robotic reproduction of general object deformation.

We present a novel technique for texture synthesis using optimization. We define a Markov Random Field (MRF)-based similarity metric for measuring the quality of synthesized texture with respect to a given input sample. This allows us to formulate the synthesis problem as minimization of an energy function, which is optimized using an Expectation Maximization (EM)-like algorithm. In contrast to most example-based techniques that do region-growing, ours is a joint optimization approach that progressively refines the entire texture. Additionally, our approach is ideally suited to allow for controllable synthesis of textures. Specifically, we demonstrate controllability by animating image textures using flow fields. We allow for general two-dimensional flow fields that may dynamically change over time. Applications of this technique include dynamic texturing of fluid animations and texture-based flow visualization.

To form a hybrid approach for image segmentation, several researches have been done to combine some techniques for better improvements. This article is concerned with image segmentation using combined methods. To separate foreground from background in image the pixel intensities have been considered. For image segmentation region growing with seed pixel is one of the most important segmentation methods. In single seeded region growing, it is very difficult to find out the proper position of the pixel during the selection. By considering the limitation of single seeded region growing an improved algorithm for region growing has proposed. The position of the seed pixel can be chosen before growing the region for segmentation using the proposed technique. Then combine this method with existing single seeded region growing algorithm. After the comparison using segmentation evaluation parameters it can be seen that, this combined method works better than others existing methods.

In this paper we give an overview of both classical and more modern morphological techniques. We will demonstrate their utility through a range of practical examples. After discussing the fundamental morphological ideas, we show how the classic morphological opening and closing filters lead to measures of size via granulometries, and we will discuss briefly their implementation. We also present an overview of morphological segmentation techniques, and the use of connected openings and thinnings will be demonstrated. This then leads us into the more recent set-theoretic notions of graph based approaches to image analysis.

Identification of human based on iris has gained increased attention in recent years. The paper focuses on novel and efficient approach of partial iris based recognition of human using pupil circle region growing and binary integrated edge intensity curve which defeats the difficulties of eyelids occlusions. The experimental results are obtained on CASIA database version-1 and show good performance with

Industrial Geometry aims at unifying existing and developing new methods and algorithms for a variety of application areas with a strong geometric component. These include CAD, CAM, Geometric Modelling, Robotics, Computer Vision and Image Processing, Computer Graphics and Scientific Visualization. In this paper, Industrial Geometry is illustrated via the fruitful interplay of the areas indicated above in the context of novel solutions of CAD related, geometric optimization problems involving distance functions: approximation with general B-spline curves and surfaces or with subdivision surfaces, approximation with special surfaces for applications in architecture or manufacturing, approximate conversion from implicit to parametric (NURBS) representation, and registration problems for industrial inspection and 3D model generation from measurement data. Moreover, we describe a 'feature sensitive' metric on surfaces, whose definition relies on the concept of an image manifold, introduced into Computer Vision and Image Processing by Kimmel, Malladi and Sochen. This metric is sensitive to features such as smoothed edges, which are characterized by a significant deviation of the two principal curvatures. We illustrate its applications at hand of feature sensitive curve design on surfaces and local neighborhood definition and region growing as an aid in the segmentation process for reverse engineering of geometric objects.

We propose a novel algorithm for unsupervised segmentation of color images. The proposed approach utilizes a dynamic color gradient thresholding scheme that guides the region growing process. Given a color image, a weighted vectorbased color gradient map is generated. Seeds are identified and a dynamic threshold is then used to perform reliable growing of regions on the weighted gradient map. Over-segmentation, if any, is addressed by a Similarity Measurebased region merging stage to produce the final segmented image. Comparative results demonstrate the effectiveness of this algorithm for color image segmentation.

This work explores the use of characterization features extracted based on breast-mass contours obtained by automated segmentation methods, for the classification of masses in mammograms according to their diagnosis (benign or malignant). Two sets of mass contours were obtained via two segmentation methods (a dynamic-programming-based method and a constrained region-growing method), and simplified versions of these contours (modeling the contours as ellipses) were employed to extract a set of six features designed for characterization of mass margins (contrast between foreground region and background region, coefficient of variation of edge strength, two measures of the fuzziness of mass margins, a measure of spiculation based on relative gradient orientation, and a measure of spiculation based on edge-signature information). Three popular classifiers (Bayesian classifier, Fisher's linear discriminant, and a support vector machine) were then used to predict the diagnosis of a set of 349 masses based on each of said features and some combinations of these. The systems (each system consists of a segmentation method, a featureset, and a classifier) were compared with each other in terms of their performance on the diagnosis of the set of breast masses. It was found that, although there was a percent difference of about 14% in the average segmentation quality between methods, this was translated into an average percent difference of only 4% in the classification performance. It was also observed that the spiculation feature based on edgesignature information was distinctly better than the rest of the features, although it is not very robust to changes in the quality of the segmentation. All systems were more efficient in predicting the diagnosis of benign masses than that of the malignant masses, resulting in low sensitivity and high specificity values (e.g. 0.6 and 0.8, respectively) since the positive class in the classification experiments is the set of malignant masses. It was concluded that features extracted from automated contours can contribute to the diagnosis of breast masses in screening programs by correctly identifying a majority of benign masses.

The object-based methodology is one of the most commonly used strategies for processing high spatial resolution images. A prerequisite to object-based image analysis is image segmentation, which is normally defined as the subdivision of an image into separated regions. This study proposes a new image segmentation methodology based on a self-calibrating multi-band region growing approach. Two multispectral aerial images were used in this study. The unsupervised image segmentation approach begins with a first step based on a bidirectional filtering, in order to eliminate noise, smooth the initial image and preserve edges. The results are compared with ones obtained from Definiens Developper software.

Objective The aim of this study was to evaluate a 3D tumor segmentation method for fluorodeoxyglucose positron emission tomography (FDG-PET) in the context of noninvasive estimation of tumor metabolic length (L m), as it correlates with surgical pathology and phantom results. Methods Thirty-four patients (7 women, 27 men) with esophageal cancer were retrospectively evaluated. All patients underwent FDG-PET-computed tomography (CT) imaging following endoscopic ultrasound (EUS). Seventeen patients had esophagectomy after PET/CT, without prior neoadjuvant therapy. Tumor length was assessed by EUS (L e, n = 31) and histopathology (L p, n = 17). Images were evaluated quantitatively with a 3D threshold-based region-growing program (Medical Image Processing Analysis and Visualization). L m, total metabolic volume (V m), maximum standardized uptake value (SUVmax), and average SUV (SUVa) over the entire tumor were obtained for several threshold values (mean activity in the liver plus 0-, 1-, 2-, 3-, and 4-SD of the activity in the liver). Results L m showed a good correlation with L p for all thresholds (best correlation for L m(2-SD), r = 0.74, P < 0.001). A positive nonsignificant correlation was observed between L p and L e (r = 0.30, P = 0.29). L m(2-SD) correlated well with L e (r = 0.71, P < 0.001). Good correlations were also observed between L m(2-SD) and V m(2-SD) (r = 0.89, P < 0.001) and SUVa(2-SD) (r = 0.38, P < 0.05). V m(2-SD) also had a significant correlation with L p (r = 0.61, P < 0.05) and L e (r = 0.57, P < 0.001). Conclusions FDG-PET-derived tumor metabolic length of untreated esophageal carcinomas correlates well with surgical pathology results, and provides preliminary evidence that noninvasive delineation of the superior and inferior extent of viable tumor involvement might be feasible using computer-generated metabolic length measurements.

In this paper, we propose an image-based system for Arabic Sign Language (ArSL) recognition. The algorithm starts by detecting the face of the signer using a Gaussian skin color model. The centroid of the detected face is then used as a reference point for tracking the hands' movements. The hands regions are segmented using a region growing algorithm assuming the signer wears a yellow and an orange colored gloves. From the segmented hands regions, an optimal set of features is extracted. To represent the time varying feature patterns, a Hidden Markov Model (HMM) is then used. Before using HMM in testing, the number of states and the number of Gaussian mixtures are optimized. The proposed system was implemented for both signer dependent and signer independent conditions. The experimental results show that an accuracy of more than 95% can be achieved with a large database of 300 signs. The results outperform previous work on ArSL mainly restricted to small vocabulary size.

A Novel Wavelet Transformation-Based Detection of Masses in digital mammograms (WTBDM) is proposed in this paper that enables for the early prognosis of breast cancer. The wavelet analysis is explored for analyzing and identifying strong variations in intensities within the mammographic data which highlights and recognizes the masses effectively. The proposed algorithm, in addition to wavelet transformation, uses morphological preprocessing, region properties and seeded region growing to remove the digitization noises, to remove the pectoral muscle and to suppress radiopaque artifacts, thus segmenting the abnormal masses accurately. The combined potential of wavelet and region growing helps for effective mass segmentation that vouches the merit of the proposed technique.

In this paper, the real-time segmentation of surgical instruments with color images used in minimally invasive surgery is addressed. This work has been developed in the scope of the robotized laparoscopic surgery, specifically for the detection and tracking of gray regions and accounting for images of metallic instruments inside the abdominal cavity. With this environment, the moving background due to the breathing motion, the non-uniform and time-varying lighting conditions and the presence of specularities are the main difficulties to overcome. Then, to achieve an automatic color segmentation suitable for robot control, we developed a technique based on a discriminant color feature with robustness capabilities with respect to intensity variations and specularities. We also designed an adaptive region growing with automatic region seed detection and a model-based region classification, both dedicated to laparoscopy. The foreseen application is a good training ground to evaluate the proposed technique and the effectiveness of this work has been demonstrated through experimental results with endoscopic image sequences to efficiently locate boundaries of a landmark-free needle-holder at half the video-rate.

The individualization of an object from a digital image is a common problem in the field of image processing. We propose a modified version of the watershed algorithm for image segmentation. We have presented an adaptive masking and a threshoding mechanism over each color channel to overcome over segmentation problem, before combining the segmentation from each channel into the final one. Our proposed method ensures accuracy and quality of the 10 different kinds of color images. The experimental results are obtained using image quality assessment (IQA) metrics such as PSNR, MSE, PSNR RGB and Color Image Quality Measure (CQM) based on reversible YUV color transformation. Consequently, the proposed modified watershed approach can enhance the image segmentation performance. Similarly, it is worth noticing that our proposed MWS approach is faster than many other segmentation algorithms, which makes it appropriate for real-time application. According to the visual and quantitative verification, the proposed algorithm is performing better than three other algorithms.

A new algorithm is proposed for triangular mesh reconstruction from 3D scattered points based on the existing intrinsic property driven (IPD) method. The improvements include a new approach to determine the seed triangle, a new approach to define the influence region for active edge, and a new approach to select the best active point. The new triangle is also tested with the constraint of geometric integrity. Our algorithm has been tested on some unorganized 3D point clouds. From the experimental results it can be found that our approach has the ability to generate more accurate details in the recovered surfaces.

This work explores the use of characterization features extracted based on breast-mass contours obtained by automated segmentation methods, for the classification of masses in mammograms according to their diagnosis (benign or malignant). Two sets of mass contours were obtained via two segmentation methods (a dynamic-programming-based method and a constrained region-growing method), and simplified versions of these contours (modeling the contours

Quantitative synchrotron micro-CT makes it possible to visualize remodeling zones having different mineral concentrations within bone tissue. However, so far their segmentation has only been performed by simple thresholding which is insufficient due to noise and weak contrast. In this paper, we propose a new customized region growing approach to segment each remodeling zone taking into account some biological prior. A constrained region growing strategy based on distance map and watershed is developed to control the shape of the detected regions. The statistical analysis of the results on synthetic noisy images and the application to experimental micro-CT data show the efficiency of the method. ᭧

The brain tumors are the mass of undifferentiated cells which form uncontrolled proliferation of cells in the brain. In this paper, I projected segmentation of brain tumor from Magnetic Resonance Images (MRI) using the Region Growing method to track tumor objects in brain images. The key concept in this technique is to separate the position of tumor objects exactly from other items of MR images. Region Growing is useful and simple method which uses several criteria to measure the characteristics of pixel and its neighborhood and also avoids characteristic segmentation errors. Several other algorithms have been proposed to segment brain tumor. Experiments show that the method can successfully achieve segmentation for MR brain images to help pathologists distinguish exactly lesion size and region.

Breast cancer is the second most common type of cancer and also one of the leading cause of cancer death worldwide. Mammography, which uses X-ray technology to image the breast, is currently the method of choice of early detection of breast cancer. This paper introduces a method for detecting suspicious masses on mammograms, which are efficient applicable for circumstances in Vietnam. The method consists of three steps: background subtraction, templace matching and region growing. The method has been tested on the Mammography Image Analysis Society (MIAS) database. A relatively high detection rate and low false positive per image implies that the method can be a quite effective tool in diagnosing breast cancer.

We are developing a computer-aided detection (CAD) system for breast masses on full field digital mammographic (FFDM) images. To develop a CAD system that is independent of the FFDM manufacturer's proprietary preprocessing methods, we used the raw FFDM image as input and developed a multiresolution preprocessing scheme for image enhancement. A two-stage prescreening method that combines gradient field analysis with gray level information was developed to identify mass candidates on the processed images. The suspicious structure in each identified region was extracted by clustering-based region growing. Morphological and spatial gray-level dependence texture features were extracted for each suspicious object. Stepwise linear discriminant analysis (LDA) with simplex optimization was used to select the most useful features. Finally, rule-based and LDA classifiers were designed to differentiate masses from normal tissues. Two data sets were collected: a mass data set containing 110 cases of two-view mammograms with a total of 220 images, and a no-mass data set containing 90 cases of two-view mammograms with a total of 180 images. All cases were acquired with a GE Senographe 2000D FFDM system. The true locations of the masses were identified by an experienced radiologist. Free-response receiver operating characteristic analysis was used to evaluate the performance of the CAD system. It was found that our CAD system achieved a case-based sensitivity of 70%, 80%, and 90% at 0.72, 1.08, and 1.82 false positive (FP) marks/image on the mass data set. The FP rates on the no-mass data set were 0.85, 1.31, and 2.14 FP marks/image, respectively, at the corresponding sensitivities. This study demonstrated the usefulness of our CAD techniques for automated detection of masses on FFDM images.

In the last decade, airborne laser scanning (ALS) systems have become an alternative source for the acquisition of altimeter data. Compared to high resolution orthoimages, one of the main advantages of ALS is the ability of the laser beam to penetrate vegetation and reach the ground underneath. Therefore, 3D point clouds are essential data for computing Digital Terrain Models (DTM) in natural and vegetated areas. DTMs are a key product for many applications such as tree detection, flood modelling, archeology or road detection. Indeed, in forested areas, traditional image-based algorithms for road and pathway detection would partially fail due to their occlusion by the canopy cover. Thus, crucial information for forest management and fire prevention such as road width and slope would be misevaluated. This paper deals with road and pathway detection in a complex forested mountaneous area and with their geometrical parameter extraction using lidar data. Firstly, a three-step image-based methodology is proposed to detect road regions. Lidar feature orthoimages are first generated. Then, road seeds are both automatically and semi-automatically detected. And, a region growing algorithm is carried out to retrieve the full pathways from the seeds previously detected. Secondly, these pathways are vectorized using morphological tools, smoothed, and discretized. Finally, 1D sections within the lidar point cloud are successively generated for each point of the pathways to estimate more accurately road widths in 3D. We also retrieve a precise location of the pathway borders and centers, exported as vector data.