Unsupervised Segmentation Method Research Articles

Segmentation for brain magnetic resonance images using dual‐tree complex wavelet transform and spatial constrained self‐organizing tree map

ABSTRACTIn brain MR images, the noise and low‐contrast significantly deteriorate the segmentation results. In this paper, we introduce a novel application of dual‐tree complex wavelet transform (DT‐CWT), and propose an automatic unsupervised segmentation method integrating DT‐CWT with self‐organizing map for brain MR images. First, a multidimensional feature vector is constructed based on the intensity, low‐frequency subband of DT‐CWT, and spatial position information. Then, a spatial constrained self‐organizing tree map (SCSOTM) is presented as the segmentation system. It adaptively captures the complicated spatial layout of the individual tissues, and overcomes the problem of overlapping gray‐scale intensities for different tissues. SCSOTM applies a dual‐thresholding method for automatic growing of the tree map, which uses the information from the high‐frequency subbands of DT‐CWT. The proposed method is validated by extensive experiments using both simulated and real T1‐weighted MR images, and compared with the state‐of‐the‐art algorithms. © 2014 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 24, 208–214, 2014

Image segmentation by fusion of low level and domain specific information via Markov Random Fields

We propose a new segmentation method by fusing a set of top-down and bottom-up segmentation maps under the Markov Random Fields (MRF) framework. The bottom-up segmentation maps are obtained by varying the parameters of an unsupervised segmentation method, such as Mean Shift. The top-down segmentation maps are constructed from some priori information, called domain specific information (DSI), received from a domain expert in the form of general properties about the image dataset. The properties are then used to generate domain specific maps (DSM) using logical predicates. Information gathered from the fusion of the bottom-up segmentation maps together with the domain specific maps are utilized to determine the pairwise potentials in the energy function of an unsupervised MRF model. Due to the inclusion of domain specific information, this approach can be considered as a first step to semantic image segmentation under an unsupervised MRF model. The proposed system is compared with the state of the art unsupervised image segmentation methods and satisfactory results are observed.

Automatic image segmentation using salient key point extraction and star shape prior

In this paper, a new unsupervised segmentation method is proposed. The method integrates the star shape prior of the image object with salient point detection algorithm. In the proposed method, the Harris salient point detection is first applied to the color image to obtain the initial salient points. A regional contrast based saliency extraction method is then used to select rough object regions in the image. To restrict the distribution of salient points, an adaptive threshold segmentation is applied to the saliency map to get the saliency mask. And then the salient region points can be obtained by placing the saliency mask on the initial Harris salient points. In order to make sure the salient points which we get are inside the image object thus the star shape constraint can be applied to the graph cuts segmentation, the Affinity Propagation (AP) clustering is employed to find the salient key points among the salient region points. Finally, these salient key points are regarded as foreground seeds and the star shape prior is introduced to graph cuts segmentation framework to extract the foreground object. Extensive experiments and comparisons on public database are provided to demonstrate the good performance of the proposed method.

An efficient algorithm for exact evaluation of stochastic watersheds

The stochastic watershed is a method for unsupervised image segmentation proposed by Angulo and Jeulin (2007). The method first computes a probability density function (PDF), assigning to each piece of contour in the image the probability to appear as a segmentation boundary in seeded watershed segmentation with randomly selected seeds. Contours that appear with high probability are assumed to be more important. This PDF is then post-processed to obtain a final segmentation. The main computational hurdle with the stochastic watershed method is the calculation of the PDF. In the original publication by Angulo and Jeulin, the PDF was estimated by Monte Carlo simulation, i.e., repeatedly selecting random markers and performing seeded watershed segmentation. Meyer and Stawiaski (2010) showed that the PDF can be calculated exactly, without performing any Monte Carlo simulations, but do not provide any implementation details. In a naive implementation, the computational cost of their method is too high to make it useful in practice. Here, we extend the work of Meyer and Stawiaski by presenting an efficient (quasi-linear) algorithm for exact computation of the PDF. We demonstrate that in practice, the proposed method is faster than any previously reported method by more than two orders of magnitude. The algorithm is formulated for general undirected graphs, and thus trivially generalizes to images with any number of dimensions.

Algorithms for automatic segmentation of bovine embryos produced in vitro

In vitro production has been employed in bovine embryos and quantification of lipids is fundamental to understand the metabolism of these embryos. This paper presents a unsupervised segmentation method for histological images of bovine embryos. In this method, the anisotropic filter was used in the differents RGB components. After pre-processing step, the thresholding technique based on maximum entropy was applied to separate lipid droplets in the histological slides in different stages: early cleavage, morula and blastocyst. In the postprocessing step, false positives are removed using the connected components technique that identify regions with excess of dye near pellucid zone. The proposed segmentation method was applied in 30 histological images of bovine embryos. Experiments were performed with the images and statistical measures of sensitivity, specificity and accuracy were calculated based on reference images (gold standard). The value of accuracy of the proposed method was 96% with standard deviation of 3%.

Sub-scene segmentation using constraints based on Gestalt principles

In this paper, an unsupervised sub-scene segmentation method is proposed. It emphasizes on generating more integrated and semantically consistent regions instead of homogeneous but detailed over-segmented regions usually produced by conventional segmentation methods. Several properties of sub-scenes are explored such as proximity grouping, area of influence, similarity and harmony based on psychological principles. These properties are formulated into constraints that are used directly in the proposed sub-scene segmentation. A self-determined approach is conducted to get the optimal segmentation result based on the characteristics of each image in an unsupervised manner. The proposed method is evaluated over three datasets. For quantitative evaluation, the performance of the proposed method is on par with state-of-the-art unsupervised segmentation methods; for qualitative evaluation, the proposed method handles various sub-scenes well, and produces neater results. The sub-scenes segmented by the proposed method are generally consistent with natural scene categories.

Unsupervised tissue segmentation from dynamic contrast-enhanced magnetic resonance imaging

Design, implement, and validate an unsupervised method for tissue segmentation from dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). For each DCE-MRI acquisition, after a spatial registration phase, the time-varying intensity of each voxel is represented as a sparse linear combination of adaptive basis signals. Both the basis signals and the sparse coefficients are learned by minimizing a functional consisting of a data fidelity term and a sparsity inducing penalty. Tissue segmentation is then obtained by applying a standard clustering algorithm to the computed representation. Quantitative estimates on two real data sets are presented. In the first case, the overlap with expert annotation measured with the DICE metric is nearly 90% and thus 5% more accurate than state-of-the-art techniques. In the second case, assessment of the correlation between quantitative scores, obtained by the proposed method against imagery manually annotated by two experts, achieved a Pearson coefficient of 0.83 and 0.87, and a Spearman coefficient of 0.83 and 0.71, respectively. The sparse representation of DCE MRI signals obtained by means of adaptive dictionary learning techniques appears to be well-suited for unsupervised tissue segmentation and applicable to different clinical contexts with little effort.

Unsupervised multiphase color–texture image segmentation based on variational formulation and multilayer graph

This paper proposes an unsupervised variational segmentation approach of color–texture images. To improve the description ability, the compact multi-scale structure tensor, total variation flow, and color information are integrated to extract color–texture information. Since heterogeneous image object and nonlinear variation exist in color–texture image, it is not appropriate to use one single/multiple constant in the Chan and Vese (CV) model to describe each phase [1,2]. Therefore, a multiphase successive active contour model (MSACM) based on the multivariable Gaussian distribution is presented to describe each phase. As geodesic active contour (GAC) has a stronger ability in capturing boundary. To inherit the advantages of edge-based model and region-based model, we incorporate the GAC into the MSACM to enhance the detection ability for concave edge. Although multiphase optimization of our proposed MSACM is a NP hard problem, we can discretely and approximately solve it by a multilayer graph method. In addition, to segment the color–texture image automatically, an adaptive iteration convergence criterion is designed by incorporating the local Kullback–Leibler distance and global phase label, so that we can control the segmentation process converges. Comparing to state-of-the-art unsupervised segmentation methods on a substantial of color texture images, our approach achieves a significantly better performance on capture ability of homogeneous region/smooth boundary and accuracy.

New method for unsupervised segmentation of moving objects in infrared videos

A new method for unsupervised segmentation of moving objects in infrared videos is presented. This method consists of two steps: difference image quantization and spatial segmentation. In the first step, the changed pixels in the difference image are quantized to several classes by using Bayes decision. It can be used to cluster the changed pixels belonging to the same moving object together. The pixels of the difference image are replaced by their corresponding class labels, thus forming a class-map of the difference image. In the second step, each class in the class-map is considered as a subset of the possible seeds of moving objects. A self-adaptive region growing method is then used to image segmentation on the basis of these different subsets. One of the focuses of this work is on spatial segmentation, where a criterion is proposed for evaluation of moving object segmentation without ground truth in infrared videos. This criterion is used to evaluate the performance of the segmentation masks grown from different subsets of the possible seeds. The best segmented image is determined to be the final segmentation result. Experiments show the advantage and robustness of the proposed algorithm on real infrared videos.

A Comparative Study of Color Image Segmentation Using Hard, Fuzzy,Rough Set Based Clustering Techniques

Image segmentation is the process of subdividing an image into its constituent parts and extracting these parts of interest, which are the objects. Colour image segmentation emerges as a new area of research. It can solve many contemporary problems in medical imaging, mining and mineral imaging, bioinformatics, and material sciences. Naturally, color image segmentation demands well defined borders of different objects in an image. So, there is a fundamental demand of accuracy. The segmented regions or components should not be further away from the true object than one or a few pixels. So, there is a need for improved image segmentation technique that can segment different components precisely. Image data may have corrupted values due to the usual limitations or artifacts of imaging devices. Noisy data, data sparsity, and high dimensionality of data create difficulties in image pixel clustering. As a result, image pixel clustering becomes a harder problem than other form of data. Taking into account all the above considerations we propose an unsupervised image segmentation method using Rough-Fuzzy C-Mean a hybrid model for segmenting RGB image by reducing cluster centers using rough sets and Fuzzy C-Means Method, and also compare the effectiveness of the clustering methods such as Hard C Means (HCM), Fuzzy C Means (FCM), Fuzzy K Means (FKM), Rough C Means (RCM) with cluster validity index such as DB Index, XB Index and Dunn Index. A good clustering procedure should make the value of DB index as low as possible, for Dunn Index high value, and for XB Index low value.

An Adaptive Growing Self-Organizing Tree Map for Brain MR Image Segmentation

Segmentation of brain tissues from magnetic resonance (MR) images plays a crucial role in medical image processing. In this paper, we propose an automatic unsupervised segmentation method integrating wavelet transform with self-organizing map for brain MR image. Firstly, a multi-dimensional feature vector is constructed based on the intensity, the low-frequency subband of wavelet transform and spatial position information. Then, an adaptive growing self-organizing tree map (AGSOTM) is presented, which adaptively captures the complicated spatial layout of the individual tissues, and overcomes the problem of overlapping grey-scale intensities for different tissues. The proposed method is validated by extensive experiments using both simulated and real T1-weighted MR images, and compared with other algorithms.

Unsupervised mitochondria segmentation using recursive spectral clustering and adaptive similarity models

The unsupervised segmentation method proposed in the current study follows the evolutional ability of human vision to extrapolate significant structures in an image. In this work we adopt the perceptual grouping strategy by selecting the spectral clustering framework, which is known to capture perceptual organization features, as well as by developing similarity models according to Gestaltic laws of visual segregation. Our proposed framework applies but is not limited to the detection of cells and organelles in microscopic images and attempts to provide an effective alternative to presently dominating manual segmentation and tissue classification practice. The main theoretical contribution of our work resides in the formulation of robust similarity models which automatically adapt to the statistical structure of the biological domain and return optimal performance in pixel classification tasks under the wide variety of distributional assumptions.

Improving MR brain image segmentation using self-organising maps and entropy-gradient clustering

The primary brain image segmentation goal is to partition a given brain image into different regions representing anatomical structures. Magnetic resonance image (MRI) segmentation is especially interesting because the accurate representation of white matter, grey matter and cerebrospinal fluid provides a way to identify many brain disorders, such as dementia, schizophrenia or Alzheimer’s disease (AD). This paper presents a fully unsupervised method for MRI segmentation based on Self-Organising Maps (SOMs) and Genetic Algorithms (GAs). In particular, the proposed method is based on five stages consisting of image acquisition, feature extraction, feature selection using evolutionary computation, voxel classification using SOM, and sharp map clustering. Moreover, a novel SOM clustering mechanism is presented that defines cluster borders by considering the output space and the relationship with the input space. This clustering mechanism accommodates spatial relationships using an entropy-gradient (EG) function calculation to group the units on the output space (SOM). The complete procedure does not use any a priori knowledge regarding voxel class assignment, but reveals a fully unsupervised, automated method for MRI segmentation to directly identify different tissue classes. Our algorithm was tested using images from the Internet Brain Image Repository (IBSR), outperforming the CGMM method in WM and CSF delineation, and was found to be the most effective among other techniques. Furthermore, this method provided excellent results using high-resolution MR images provided by the Nuclear Medicine Service of the “Virgen de las Nieves” Hospital (Granada, Spain).

An unsupervised method for dominant colour region segmentation in yarn‐dyed fabrics

This paper presents a novel unsupervised approach to detect dominant colour regions standing out conspicuously in yarn-dyed fabric images. For a dominant colour region of a yarn-dyed fabric, measured by an imaging system, its individual yarn has an irregular three-dimensional shape resulting in significant colour difference among pixels of the yarn. This difference leads to difficulty in segmenting yarns into dominant colour regions. A probabilistic model is proposed in this study to associate the colour of a dominant colour region with the colours of its yarns. Based on this model, the colour histograms of a dominant colour region are first estimated from those of yarns in a yarn-dyed fabric image. Then, a hierarchical segmentation structure is devised to detect dominant colour regions in the image. Experimental results show that the proposed approach achieves satisfactory performance for dominant colour region segmentation in yarn-dyed fabric images, with high computational efficiency.

Unsupervised optimal phoneme segmentation: theory and experimental evaluation

Automatic phoneme segmentation of a speech sequence is a basic problem in speech engineering. This study investigates unsupervised phoneme segmentation without using prior information on linguistic contents and acoustic models of an input sequence. The authors formulate the unsupervised segmentation as an optimal problem by means of maximum likelihood, and show that the optimal segmentation corresponds to minimising the coding length of the input sequence. Under different assumptions, five different objective functions are developed, namely log determinant, rate distortion (RD), Bayesian log determinant, Mahalanobis distance and Euclidean distance objectives. The authors prove that the optimal segmentations have the transformation-invariant properties, introduce a time-constrained agglomerative clustering algorithm to find the optimal segmentations, and propose an efficient implementation of the algorithm by using integration functions. The experiments are carried out on the TIMIT database to compare the above five objective functions. The results show that RD achieves the best performance, and the proposed method outperforms the previous unsupervised segmentation methods.

Unsupervised segmentation method for cuboidal cell nuclei in histological prostate images based on minimum cross entropy

This paper presents a novel segmentation method for cuboidal cell nuclei in images of prostate tissue stained with hematoxylin and eosin. The proposed method allows segmenting normal, hyperplastic and cancerous prostate images in three steps: pre-processing, segmentation of cuboidal cell nuclei and post-processing. The pre-processing step consists of applying contrast stretching to the red (R) channel to highlight the contrast of cuboidal cell nuclei. The aim of the second step is to apply global thresholding based on minimum cross entropy to generate a binary image with candidate regions for cuboidal cell nuclei. In the post-processing step, false positives are removed using the connected component method. The proposed segmentation method was applied to an image bank with 105 samples and measures of sensitivity, specificity and accuracy were compared with those provided by other segmentation approaches available in the specialized literature. The results are promising and demonstrate that the proposed method allows the segmentation of cuboidal cell nuclei with a mean accuracy of 97%.

Markov Random Field and Max Entropy Criterion-based MRI Unsupervised Segmentation Method

Markov Random Field and Max Entropy Criterion-based MRI Unsupervised Segmentation Method

Unsupervised Natural Image Segmentation via Bayesian Ying–Yang Harmony Learning Theory

An unsupervised image segmentation method for natural images is proposed in this paper. We assume that texture features in natural images are distributed as a mixture of Gaussians. In order to cluster the extracted feature vectors, we modify a clustering algorithm based on Bayesian Ying–Yang (BYY) harmony learning theory with Dirichlet-Normal-Wishart prior. This algorithm can determine the number of components automatically during the clustering procedure, as long as we give a large enough initial component number. Our works in this paper have presented a complete pipeline of clustering-based image segmentation including feature extraction, robust feature clustering and methodological effective post processing. The experiments reported in this paper demonstrate that the proposed method is efficient (in terms of visual evaluation and quantitative performance measures) and performs competitively compared to the existing state-of-the-art segmentation methods.

LVQ-SVM based CAD tool applied to structural MRI for the diagnosis of the Alzheimer’s disease

This paper presents a novel computer-aided diagnosis (CAD) tool for the diagnosis of the Alzheimer’s disease (AD) using structural Magnetic Resonance Images (MRIs). The proposed method uses information learnt from the tissue distribution of Gray Matter (GM) and White Matter (WM) in the brain, which is previously obtained by an unsupervised segmentation method. The tissue distribution of control (normal) and AD images is modelled by means of Learning Vector Quantization (LVQ) algorithm, generating a set of representative prototypes of each class. The devised method projects new images onto the model vectors space for further classification using Support Vector Machine (SVM). The tool proposed here yields classification results over 90% (accuracy) for controls (normal) and Alzheimer’s disease (AD) patients and sensitivity up to 95% to AD. Moreover, statistical significance tests have been also performed in order to validate the proposed approach.

A Textural–Contextual Model for Unsupervised Segmentation of Multipolarization Synthetic Aperture Radar Images

This paper proposes a novel unsupervised, non-Gaussian, and contextual segmentation method that combines an advanced statistical distribution with spatial contextual information for multilook polarimetric synthetic aperture radar (PolSAR) data. This extends on previous studies that have shown the added value of both non-Gaussian modeling and contextual smoothing individually or for intensity channels only. The method is based on a Markov random field (MRF) model that integrates a <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">K</i> -Wishart distribution for the PolSAR data statistics conditioned to each image cluster and a Potts model for the spatial context. Specifically, the proposed algorithm is constructed based upon the stochastic expectation maximization (SEM) algorithm. A new formulation of SEM is developed to jointly perform clustering of the data and parameter estimation of the <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">K</i> -Wishart distribution and the MRF model. Experiments on simulated and real PolSAR data demonstrate the added value of using an appropriate statistical representation, in combination with contextual smoothing.