Region-based Segmentation Techniques Research Articles

A New Region-Based Minimal Path Selection Algorithm for Crack Detection and Ground Truth Labeling Exploiting Gabor Filters

Cracks are fractures or breaks that occur in materials such as concrete, metals, rocks, and other solids. Various methods are used to detect and monitor cracks; among many of them, image-based methodologies allow fast identification of the distress and easy quantification of the percentage of cracks in the scene. Two main categories can be identified: classical and deep learning approaches. In the last decade, the tendency has moved towards the use of the latter. Even though they have proven their outstanding predicting performance, they suffer some drawbacks: a “black-box” nature leaves the user blind and without the possibility of modifying any parameters, a huge amount of labeled data is generally needed, a process that requires expert judgment is always required, and, finally, they tend to be time-consuming. Accordingly, the present study details the methodology for a new algorithm for crack segmentation based on the theory of minimal path selection combined with a region-based approach obtained through the segmentation of texture features extracted using Gabor filters. A pre-processing step is described, enabling the equalization of brightness and shadows, which results in better detection of local minima. These local minimal are constrained by a minimum distance between adjacent points, enabling a better coverage of the cracks. Afterward, a region-based segmentation technique is introduced to determine two areas that are used to determine threshold values used for rejection. This step is critical to generalize the algorithm to images presenting close-up scenes or wide cracks. Finally, a geometrical thresholding step is presented, allowing the exclusion of rounded areas and small isolated cracks. The results showed a very competitive F1-score (0.839), close to state-of-the-art values achieved with deep learning techniques. The main advantage of this approach is the transparency of the workflow, contrary to what happens with deep learning frameworks. In the proposed approach, no prior information is required; however, the statistical parameters may have to be adjusted to the particular case and requirements of the situation. The proposed algorithm results in a useful tool for researchers and practitioners needing to validate their results against some reference or needing labeled data for their models. Moreover, the current study could establish the grounds to standardize the procedure for crack segmentation with a lower human bias and faster results. The direct application of the methodology to images obtained with any low-cost sensor makes the proposed algorithm an operational support tool for authorities needing crack detection systems in order to monitor and evaluate the current state of the infrastructures, such as roads, tunnels, or bridges.

Durability of Low-Pt Loading Membrane Electrode Assembly under Dynamic Driving Conditions

In this paper, a long-term durability test for 3910 h is carried out on a low Pt-loading fuel cell stack by simulating the dynamic driving cycles. The performance of a specific membrane electrode assembly (MEA) in the stack before and after the test is analyzed, revealing the high voltage decay rate of the appliance. Various electrochemical and physical characterization techniques are used to analyze the degradation mechanism of the MEA via region-based segmentation technique. The results show that the degradation of MEA performance is mainly due to the agglomeration and loss of Pt nanoparticles in the cathode catalyst. In particular, the growth of catalyst particles occurs at the inlet and outlet of the cathode. A high loss rate of Pt particles at the cathode inlet causes their aggregation at the boundary between the catalyst layer and the membrane where the loss of Pt is initiated. Finally, the migration of the unsupported Pt particles occurs due to the gravity toward the underlying cathode micro porous layer. This study had important implications for promoting the development and commercial application of cost-efficient and long-life MEAs.

Improved grey wolf optimization algorithm (iGWO) for the detection of diabetic retinopathy using convnets and region based segmentation techniques

Diabetic Retinopathy is a retinal eye disease that affects people with Diabetes Mellitus. DM is metabolic disorder and it is caused by the high glucose level in blood. This leads to eye deficiency called Diabetic Retinopathy (DR). In this paper, we proposed iGWO algorithm for the detection and diagnosis of DR stages and Region growing technique is used for segmentation of the images. Various preprocessing techniques are followed for the better enhancement of the images. An improved Grey Wolf Optimization (iGWO) algorithm are proposed to obtain the global optimum. Convnets are utilized for DR categorization. The iGWO-FFOCNN model is compared with existing technologies like SVM-GSO (Support Vector Machines – Glowworm Swam Optimization), PSO-CNN (Particle Swarm Optimization-Convolutional Neural Networks), CNN (Convolutional Neural Networks), DCNN-EMFO (Deep Convolutional Neural Networks- Enhanced Moth Flame Optimization) and MACO-CNN (Modified Ant Colony Optimization–CNN). The APTOS DR Dataset utilized for the proposed metodology. The effort is made for identifying the DR stages by using iGWO. Finally, the results confirm that iGWO-FFOCNN technique yields good performance than compared to the existing techniques in relation to F-measure, precision, recall and also in DSC (Dice Similarity Coefficient), JSC (Jaccard Similarity Coefficient) and time period.

Framework of diacritic segmentation for Arabic handwritten document

<span lang="EN-US">In <span>recent Arabic standard language and Arabic dialectal texts, diacritics and short vowels are absent. There are some exceptions have been made for the Arabic beginner learner scripts, religious texts and as well as a significant political text. In addition, the text without diacritics is considered ambiguous due to numerous words with different diacritic marks seem identical. However, this paper we present a framework for segmenting diacritics from Arabic handwritten document by using region-based segmentation technique. Since Arabic handwritten and Mushaf Al-Quran contain many diacritical marks. Hence, the diacritics must be properly extracted from Arabic handwritten document to avoid losing some good features. Furthermore, the proposed framework is devised specifically to segment diacritics from Arabic handwritten image, thus there will be no feature extraction, feature selection, and classification processes included. Besides, we will present the methodology that is used to fulfil the objectives of this paper. The pre-processing phases will be explained and more specifically segmentation phase for segmenting diacritics which is the phase we concentrate more in this article. Lastly, we will identify the proposed technique region-based segmentation to facilitate our development throughout the experimental process.</span></span>

A Hybrid Morphological Active Contour for Natural Images

Morphological active contours for image segmentation have become popular due to their low computational complexity coupled with their accurate approximation of the partial differential equations involved in the energy minimization of the segmentation process. In this paper, a morphological active contour which mimics the energy minimization of the popular Chan-Vese Active Contour without Edges is coupled with a morphological edge-driven segmentation term to accurately segment natural images. By using morphological approximations of the energy minimization steps, the algorithm has a low computational complexity. Additionally, the coupling of the edge-based and region-based segmentation techniques allows the proposed method to be robust and accurate. We will demonstrate the accuracy and robustness of the algorithm using images from the Weizmann Segmentation Evaluation Database and report on the segmentation results using the Sorensen-Dice similarity coefficient.

Automated layerwise detection of geometrical distortions in laser powder bed fusion

In-situ layerwise imaging in laser powder bed fusion (L-PBF) has been implemented by many system developers to monitor the powder bed homogeneity. Increasing attention has been recently devoted to the possibility of using the same sensing approach to detect also in-plane and out-of-plane geometrical distortions of the part while it is being produced. To this aim, seminal works investigated the suitability of various image segmentation algorithms and assessed the accuracy of layerwise dimensional and geometrical measurements. Nevertheless, there is a lack of automated methods to identify, in-situ and in-process, geometrical defects and out-of-control deviations from the nominal geometry. This study presents a methodology that combines an active contours methodology for image segmentation with a statistical process monitoring approach suitable to deal with complex geometries that change layer by layer. The proposed approach enables a data-driven and automated alarm rule to detect the onset of geometrical distortions during the build by comparing the slice contour reconstruction with the nominal geometry in each layer. Moreover, by coupling edge-based and region-based segmentation techniques, the method is sufficiently robust to be applied to imaging and illumination setups that are already available on industrial L-PBF systems. The effectiveness of the proposed approach was tested on a real case study involving the L-PBF of complex Ti6Al4V parts that exhibited local geometrical distortions.

Diacritic segmentation technique for arabic handwritten using region-based

Arabic is a broadly utilized alphabetic composition framework on the planet, and it has 28 essential letters. The letters in order was first used to compose messages in Arabic, most prominently the Qur'an the holy book of Islam. However, Arabic language has diacritics in the word or letters which are not something extra or discretionary to the language, rather they are a vital piece of it. By changing some diacritics may change both the syntax and semantics of a word by turning a word into another. However, the current researches address the foreground image and consider the diacritics as noises or secondary images. Thus, it is not suitable for Arabic handwritten. The diacritics will be removed from the image and this will lead to losing some good features. Furthermore, to extract the diacritics, the region-based segmentation technique is used. The image will be measured based on the region properties by first finding the connected component in binary image, and then we will determine the best area range measurement in that region for each image. The proposed technique region based has been tested in nine different images with different handwritten style, and successfully extracted secondary foreground images (diacritics) for each image.

Region Based Segmentation Techniques for Digital Images

This paper discusses region based segmentation techniques for digital images. For a few applications, such as image compression or recognition, we cannot handle the entire picture straightforwardly as it is unconventional and inefficient. Due to these reasons, many algorithms related to image segmentation are proposed in the literature to divide an image prior to compression or recognition. The segmentation of an image is basically done to arrange or group the image in a few fragments (districts) as specified by the elements of an image, for instance, according to the value of pixel or frequency response. Currently, many image segmentation approaches exist and are widely used in across scientific disciplines and daily human life. The segmentation approaches can be generally categorized to segmentation based on region, segmentation based on edges, and information grouping.

An Automated Brain Tumour Boundary Detection using Region based Segmentation Technique along with SVM Classifier

We presents a fully automated method for an automated brain-tumour boundary detection using region based segmentation technique along with SVM Classifier of Magnetic Resonance Imaging (MRI).The procedure is based on artificial intelligence technique and classification of each super-pixel in MRI. A number of novel image features extraction approaches including intensity-based, texture based, fractal analysis and curvatures are calculated from each super-pixel within the entire brain area in MRI to ensure a robust classification. Brain tumor is the malignant types of cancer and their classification in earlier stage is biggest issue. While curable with early classification is useful, only extremely trained specialists are capable of accurately recognizing the cancer from skin MRI data. As expertise is in limited contribute, an automated systems capable of classifying cancer could save human lives, and also help to reduce unnecessary MRI, and reduce extra costs. On the way to achieve this goal, we proposed a Brain Tumour Detection and Classification System (BTDCS) that combines recent developments in machine learning with Support Vector Machine (SVM) structure, creating hybrid algorithm of threshold based segmentation with Maximally Stable External Regions (MSER) that are capable of segmenting accurate super-pixel region from MRI, as well as analyzing the detected area and surrounding tissue for malignant. Using threshold based segmentation technique, the foreground and background component is separated into two regions. To improve the segmentation results, MSER is used with the novel concept of region detection and feature extraction mechanism. The proposed system is evaluated using the largest publicly accessible standard BRATS 2015 dataset of MRI, containing benign and malignant images. When the evaluation parameters of proposed work is compared with a few other state-of-art methods, the proposed means attains the best performance of 98.2% concerning classification accuracy using only the MSER approach and SVM as classifier. The ultimate aim of this research is to devise an automated experimental approach that can segment the tumor boundary in a fast and efficient manner.

Analysis of skin lesions using gmm-hmrf region-based segmentation technique

Analysis of skin lesions using gmm-hmrf region-based segmentation technique

GVF snake algorithm-a parallel approach

Multicore architecture is an emerging computer technology where multiple processing element will be acting as independent processing cores by sharing a common memory. Digital image segmentation is a widely used medical imaging application to extract regions of interest. GVF Active Contour is a region based segmentation technique which extracts curved and irregular shaped regions by diffusing gradient vectors and by the influence of internal and external forces. This requires prior knowledge on the geometric position and anatomical structures to locate the specific region defined within an image domain. This process requires complex mathematical calculations which in turn results in the immense consumption of CPU processing time. This may adversely affect the overall performance efficiency of the process. With the advancements in multicore technology, this processing time delay can be reduced by adapting parallelization in the computation of GVF field to the specific region of interest which is to be segmented. OpenMP is a shared memory parallel programming construct, which could implement multicore parallelism with extensive and powerful APIs thereby supporting the functionalities required to attain parallelism. This article provides a high level overview of OpenMP, its effectiveness and ease of implementation in adapting parallelism to existing traditional sequential methods using instruction, data and loop level parallelism. Performance comparison could be done with sequential versions of the program written in Matlab, Java and C languages with the proposed parallelized version of OpenMP. The result is also comparable with different operating systems like Windows and Linux.

SEGMENTATION AND CLASSIFICATION OF CERVICAL CYTOLOGY IMAGES USING MORPHOLOGICAL AND STATISTICAL OPERATIONS

Cervical cancer that is a disease, in which malignant (cancer) cells form in the tissues of the cervix, is one of the fourth leading causes of cancer death in female community worldwide. The cervical cancer can be prevented and/or cured if it is diagnosed in the pre-cancerous lesion stage or earlier. A common physical examination technique widely used in the screening is called Papanicolaou test or Pap test which is used to detect the abnormality of the cell. Due to intricacy of the cell nature, automating of this procedure is still a herculean task for the pathologist. This paper addresses solution for the challenges in terms of a simple and novel method to segment and classify the cervical cell automatically. The primary step of this procedure is pre-processing in which de-nosing, de-correlation operation and segregation of colour components are carried out, Then, two new techniques called Morphological and Statistical Edge based segmentation and Morphological and Statistical Region Based segmentation Techniques- put forward in this paper, and that are applied on the each component of image to segment the nuclei from cervical image. Finally, all segmented colour components are combined together to make a final segmentation result. After extracting the nuclei, the morphological features are extracted from the nuclei. The performance of two techniques mentioned above outperformed than standard segmentation techniques. Besides, Morphological and Statistical Edge based segmentation is outperformed than Morphological and Statistical Region based Segmentation. Finally, the nuclei are classified based on the morphological value. The segmentation accuracy is echoed in classification accuracy. The overall segmentation accuracy is 97%.

Dictionary Learning Level Set

We propose a novel region based segmentation technique using dictionary learning. In a previous work we have developed a method which uses a set of pre-specified Legendre basis functions to perform region based segmentation of an object in presence of heterogeneous illumination. We hypothesize that in problems where a set of training images for the object is available for analysis (such as depth image sequence of blood vessels via ultrasound imaging), segmentation accuracy can be significantly improved by learning the basis functions instead of specifying them implicitly. The salient idea of this letter is to compute the optimal set of functions to model the region intensities. Our solution to this problem involves the integration of a level set segmentation methodology with the dictionary learning framework. This provides an elegant solution to deal with intensity inhomogeneities prevalent in many imaging applications such as ultrasound and fluorescence microscopy. The proposed algorithm, Dictionary Learning Level Set (DL2S) is used to segment ultrasound images of blood vessels captured using low cost, portable ultrasound devices employed in a phlebotomy application. Qualitative and quantitative results obtained from this dataset suggest efficacy of D2LS with an associated improvement in the average Dice index of 12% over the relevant competitors.

A contour-based segmentation algorithm for triangle meshes in 3D space

This paper introduces the first contour-based mesh segmentation algorithm that we may find in the literature, which is inspired in the edge-based segmentation techniques used in image analysis, as opposite to region-based segmentation techniques. Its leading idea is to firstly find the contour of each region, and then to identify and collect all of its inner triangles. The encountered mesh regions correspond to ups and downs, which do not need to be strictly convex nor strictly concave, respectively. These regions, called relaxedly convex regions (or saliences) and relaxedly concave regions (or recesses), produce segmentations that are less-sensitive to noise and, at the same time, are more intuitive from the human point of view; hence it is called human perception-oriented (HPO) segmentation. Besides, and unlike the current state-of-the-art in mesh segmentation, the existence of these relaxed regions makes the algorithm suited to both nonfreeform and freeform objects.

A Hybrid Morphological Active Contour for Natural Images

Morphological active contours for image segmentation have become popular due to their low computational complexity coupled with their accurate approximation of the partial differential equations involved in the energy minimization of the segmentation process. In this paper, a morphological active contour which mimics the energy minimization of the popular Chan-Vese Active Contour without Edges is coupled with a morphological edge-driven segmentation term to accurately segment natural images. By using morphological approximations of the energy minimization steps, the algorithm has a low computational complexity. Additionally, the coupling of the edge-based and region-based segmentation techniques allows the proposed method to be robust and accurate. We will demonstrate the accuracy and robustness of the algorithm using images from the Weizmann Segmentation Evaluation Database and report on the segmentation results using the Sorensen-Dice similarity coefficient.

Importance of Mean Shift in Remote Sensing Segmentation

The segmentation of an image can be considered as a preprocessing step for many of the algorithms such as object detection and identification. Many algorithms exist in the field of remote sensing for segmentation, in every case the desirable output of segmentation is a well defined region or features of object which can be distinguished from another. The desirable features include well based edges, gradients, textures etc. For a real world image the desirable features are difficult to identify. With the advancement of the high resolution imagery, the features which defined the object are too finely defined. It also contains noises which are due to fine edges, gradient variations and non-uniform textures. Segmentation algorithms are plenty in image processing which can be broadly categorized as edge based, color based and textured based. Edge based techniques fails due to the noise content. Textures are not uniform in real world images leading to problems of segmentation. Color based techniques need homogeneous regions which can distinguish objects, which will be difficult with the gradient variation. Defined in this paper is the importance of non-parametric clustering technique called Mean Shift, which in its inherent nature is able to cluster regions according to the desirable properties. The paper is a study on Mean-shift and its probable use in clustering of remote sensing imagery. Rather than a theoretical paper, the paper is arranged as an application based survey which can show the possible use and importance of mean shift in remote sensing. Segmentation is an image processing technique which is nearly an unavoidable preprocessing step. There are many techniques used for segmentation in the field of IP, which can be adapted to remote sensing applications. Adapting the algorithm to remote sensing is same as using the techniques for real world datasets. For a remote sensing system the desirable output of segmentation are well defined regions or features of object which can be distinguished from another. These features are affected by the homogeneity of regions. Homogeneity can be recognized as same color or same texture for a region. The remote sensing images of a natural region such as forest have regions of greenery more. This helps in the segmentation of those regions which may be a good part of the image. If it is an urban image the number of regions will be more with lots of small homogenous regions. In real world remote sensing imagery, the homogeneity may not be evident as expected. There will often be small gradient and textural variation. The images also may contain large number of regions, which may not be known prior to execution. These interfere in the segmentation steps and further processing using those regions. Some of the earlier version of segmentation included edge based, contour based, model based, template based and also region based segmentation. Schiewe (1) has stated the usage of edge based segmentation as having the problem of giving large number of edges, due to trees in natural images and blocks in urban areas. The edges were needed then to group and form meaningful geometry. The algorithms are not viable, due to the reason that edge grouping to segment a region with specific geometry is of non-polynomial complexity. Using region based segmentation techniques such as K-Means segmentation techniques (2) has the inherent disadvantage of knowing the number of regions prior to segmentation, which is not known in real world cases. Contour based techniques such as Active contour models (3), sometimes used in remote sensing also has the problem of knowing the approximate location of the region and also the noises will hinder its performance. A template based approach (4) is not scale invariant and using multiple scales is not suited in real time detection. Color based segmentation is also not suited for the problem due to the fact that the object may have gradient variation and color ranges. Schiewe (1) has also given an overview of the use of segmentation technique in remote sensing. He has stated some of the method with example and also given some of the applications in which segmentation can be used.

A comparative study of Image Region-Based Segmentation Algorithms

Image segmentation has recently become an essential step in image processing as it mainly conditions the interpretation which is done afterwards. It is still difficult to justify the accuracy of a segmentation algorithm, regardless of the nature of the treated image. In this paper we perform an objective comparison of region-based segmentation techniques such as supervised and unsupervised deterministic classification, non-parametric and parametric probabilistic classification. Eight methods among the well-known and used in the scientific community have been selected and compared. The Martin's(GCE, LCE), probabilistic Rand Index (RI), Variation of Information (VI) and Boundary Displacement Error (BDE) criteria are used to evaluate the performance of these algorithms on Magnetic Resonance (MR) brain images, synthetic MR image, and synthetic images. MR brain image are composed of the gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) and others, and the synthetic MR image composed of the same for real image and the plus edema, and the tumor. Results show that segmentation is an image dependent process and that some of the evaluated methods are well suited for a better segmentation. Keywords—Evaluation criteria; Martin's; Rand Index; Image Segmentation; Magnetic resonance image. I. INTRODUCTION

Automatic measurement of surface tension from noisy images using a component labeling method

Automatic surface tension measurement of turbid liquids using a bubble method (e.g., pendant/captive bubble) or interfacial tension measurement of some liquid–liquid systems necessitates an image analysis scheme robust against noise. Measuring the surface tension of lung surfactant–polymer systems using the combination of axisymmetric drop shape analysis (ADSA) and a captive bubble is such an example. We have recently developed a sophisticated image analysis scheme featuring the combination of the Canny edge detector and a novel edge smoothing technique called axisymmetric liquid fluid interfaces-smoothing (ALFI-S). The Canny edge detector is highly noise-resistant and ALFI-S is efficient in further removing noise close to the bubble profile (i.e., adhering noise). However, the procedure of eliminating noise far away from the bubble profile (i.e., isolated noise) is still not satisfactory in the previous scheme. Here a component labeling method is developed to automatically remove isolated noise. Component labeling refers to the process of detecting connected objects in a digital image. Once labeling is completed, information, such as area, location and pixel intensity, of these separate objects can be obtained. Hence, component labeling can be used as a region-based segmentation technique that is capable of differentiating the bubble and the isolated noise. The new component labeling-based noise reduction method is used to analyze captive bubble images with different amounts of noise. It is found that with the help of the labeling procedure smooth edges can be detected with a simple set of user-specified parameters even from images with extensive noise. Combined with the previous image analysis scheme, the component labeling method significantly improves the effectiveness and reliability of ADSA in automatically measuring surface tension from noisy images.

Edge- and region-based segmentation technique for the extraction of large, man-made objects in high-resolution satellite imagery

A new object-oriented segmentation approach with special focus on shape analysis was developed for the extraction of large, man-made objects, especially agricultural fields, in high-resolution panchromatic satellite imagery. The approach, a combination of region- and edge-based techniques, includes new methods for the evaluation of straight edges, edge preserving degradation, and edge-guided region growing.

Extracting characteristic segments in high-resolution panchromatic imagery as basic information for object-driven image analysis

For the monitoring of urban areas, high-resolution aerial image data are still primarily used to detect individual man-made structures. Image segmentation techniques extract and provide basic information such as regions of similar features or typical object edges for further model-driven analysis. Therefore, the quality of the final object recognition depends strongly on this information. In this paper, the potential of a combined edge- and region-based segmentation technique for the purpose of detecting characteristic man-made objects is tested and compared to standard techniques using a new quality measure. The presented approach proves its value, especially for objects with low-contrast boundaries.