Three processing characteristics of visual texture segmentation

Three mechanisms are outlined which are sufficient to determine texture segmentation or discrimination. They are: (1) convolution of detector profiles with the input image; (2) impletion, where the perceptual 'filling in' of the input surface occurs via a nonlinear filtering operation on each detector's output (3) grouping, where areas are segregated according to their differences in detector responses after impletion occurs. These mechanisms are compared with those proposed to occur in human visual texture discrimination.     

Chromosomal regions in prostatic carcinomas studied by comparative genomic hybridization, hierarchical cluster analysis and self-organizing feature maps.

Comparative genomic hybridization (CGH) is an established genetic method which enables a genome-wide survey of chromosomal imbalances. For each chromosome region, one obtains the information whether there is a loss or gain of genetic material, or whether there is no change at that place. Therefore, large amounts of data quickly accumulate which must be put into a logical order. Cluster analysis can be used to assign individual cases (samples) to different clusters of cases, which are similar and where each cluster may be related to a different tumour biology. Another approach consists in a clustering of chromosomal regions by rewriting the original data matrix, where the cases are written as rows and the chromosomal regions as columns, in a transposed form. In this paper we applied hierarchical cluster analysis as well as two implementations of self-organizing feature maps as classical and neuronal tools for cluster analysis of CGH data from prostatic carcinomas to such transposed data sets. Self-organizing maps are artificial neural networks with the capability to form clusters on the basis of an unsupervised learning rule. We studied a group of 48 cases of incidental carcinomas, a tumour category which has not been evaluated by CGH before. In addition we studied a group of 50 cases of pT2N0-tumours and a group of 20 pT3N0-carcinomas. The results show in all case groups three clusters of chromosomal regions, which are (i) normal or minimally affected by losses and gains, (ii) regions with many losses and few gains and (iii) regions with many gains and few losses. Moreover, for the pT2N0- and pT3N0-groups, it could be shown that the regions 6q, 8p and 13q lay all on the same cluster (associated with losses), and that the regions 9q and 20q belonged to the same cluster (associated with gains). For the incidental cancers such clear correlations could not be demonstrated.     

Self-organized formation of topologically correct feature maps

This work contains a theoretical study and computer simulations of a new self-organizing process. The principal discovery is that in a simple network of adaptive physical elements which receives signals from a primary event space, the signal representations are automatically mapped onto a set of output responses in such a way that the responses acquire the same topological order as that of the primary events. In other words, a principle has been discovered which facilitates the automatic formation of topologically correct maps of features of observable events. The basic self-organizing system is a one- or two-dimensional array of processing units resembling a network of threshold-logic units, and characterized by short-range lateral feedback between neighbouring units. Several types of computer simulations are used to demonstrate the ordering process as well as the conditions under which it fails.     

Kohonen's feature maps for fly ash categorization

Fly ash is a common admixture used in concrete and may constitute up to 50% by weight of the total binder material. Incorporation of fly ash in Portland-cement concrete is highly desirable due to technological, economic, and environmental benefits. This article demonstrates the use of artificial intelligence neural networks for the classification of fly ashes in to different groups. Kohonen's Self Organizing Feature Maps is used for the purpose. As chemical composition of fly ash is crucial in the performance of concrete, eight chemical attributes of fly ashes have been considered. The application of simple Kohonen's one-dimensional feature maps permitted to differentiate three main groups of fly ashes. Three one-dimensional feature maps of topology 8-16, 8-24 and 8-32 were explored. The overall classification result of 8-16 topology was found to be significant and encouraging. The data pertaining to 80 fly ash samples were collected from standard published works. The categorization was found to be excellent and compares well with Canadian Standard Association's [CSA A 3000] classification scheme.     

Automated segmentation of molecular subunits in electron cryomicroscopy density maps

Electron cryomicroscopy (cryoEM) is capable of imaging large macromolecular machines composed of multiple components. However, it is currently only possible to achieve moderate resolution at which it may be possible to computationally extract the individual components in the machine. In this work, we present application details of an automated method for detecting and segmenting the components of a large machine in an experimentally determined density map. This method is applicable to object with and without symmetry and takes advantage of global and local symmetry axes if present. We have applied this segmentation algorithm to several cryoEM data sets already deposited in EMDB with various complexities, symmetries and resolutions and validated the results using manually segmented density and available structures of the components in the PDB. As such, automated segmentation could become a useful tool for the analysis of the ever-increasing number of structures of macromolecular machines derived from cryoEM.     

Edged watershed segmentation: A semi-interactive algorithm for segmentation of low-resolution maps from electron cryomicroscopy

Electron cryomicroscopy (cryo-EM) allows for the structural analysis of large protein complexes that may be difficult to study by other means. Frequently, maps of complexes from cryo-EM are obtained at resolutions between 10 and 25 Å. To aid in the interpretation of these medium- to low-resolution maps, they may be subdivided into three-dimensional segments representing subunits or subcomplexes. This division is often accomplished using a manual segmentation approach. While extremely useful, manual segmentation is subjective. We have developed a novel semi-interactive segmentation algorithm that can incorporate prior knowledge of subunit composition or structure without biasing the boundaries between subunits or subcomplexes. This algorithm has been characterized with experimental and simulated cryo-EM density maps at resolutions between 10 and 25 Å.     

Nhs: network-based hierarchical segmentation for cryo-electron microscopy density maps

Cryo-electron microscopy (cryo-EM) experiments yield low-resolution (3-30 ?) 3D-density maps of macromolecules. These density maps are segmented to identify structurally distinct proteins, protein domains, and subunits. Such partitioning aids the inference of protein motions and guides fitting of high-resolution atomistic structures. Cryo-EM density map segmentation has traditionally required tedious and subjective manual partitioning or semisupervised computational methods, whereas validation of resulting segmentations has remained an open problem in this field. We introduce a network-based hierarchical segmentation (Nhs) method, that provides a multi-scale partitioning, reflecting local and global clustering, while requiring no user input. This approach models each map as a graph, where map voxels constitute nodes and weighted edges connect neighboring voxels. Nhs initiates Markov diffusion (or random walk) on the weighted graph. As Markov probabilities homogenize through diffusion, an intrinsic segmentation emerges. We validate the segmentations with ground-truth maps based on atomistic models. When implemented on density maps in the 2010 Cryo-EM Modeling Challenge, Nhs efficiently and objectively partitions macromolecules into structurally and functionally relevant subregions at multiple scales.     

Heterogeneity in genetic admixture across different regions of Argentina

The population of Argentina is the result of the intermixing between several groups, including Indigenous American, European and African populations. Despite the commonly held idea that the population of Argentina is of mostly European origin, multiple studies have shown that this process of admixture had an impact in the entire Argentine population. In the present study we characterized the distribution of Indigenous American, European and African ancestry among individuals from different regions of Argentina and evaluated the level of discrepancy between self-reported grandparental origin and genetic ancestry estimates. A set of 99 autosomal ancestry informative markers (AIMs) was genotyped in a sample of 441 Argentine individuals to estimate genetic ancestry. We used non-parametric tests to evaluate statistical significance. The average ancestry for the Argentine sample overall was 65% European (95%CI: 63–68%), 31% Indigenous American (28–33%) and 4% African (3–4%). We observed statistically significant differences in European ancestry across Argentine regions [Buenos Aires province (BA) 76%, 95%CI: 73–79%; Northeast (NEA) 54%, 95%CI: 49–58%; Northwest (NWA) 33%, 95%CI: 21–41%; South 54%, 95%CI: 49–59%; p<0.0001] as well as between the capital and immediate suburbs of Buenos Aires city compared to more distant suburbs [80% (95%CI: 75–86%) versus 68% (95%CI: 58–77%), p = 0.01]. European ancestry among individuals that declared all grandparents born in Europe was 91% (95%CI: 88–94%) compared to 54% (95%CI: 51–57%) among those with no European grandparents (p<0.001). Our results demonstrate the range of variation in genetic ancestry among Argentine individuals from different regions in the country, highlighting the importance of taking this variation into account in genetic association and admixture mapping studies in this population.     

Quantitative evaluation methods of skin condition based on texture feature parameters

In order to quantitatively evaluate the improvement of the skin condition after using skin care products and beauty, a quantitative evaluation method for skin surface state and texture is presented, which is convenient, fast and non-destructive. Human skin images were collected by image sensors. Firstly, the median filter of the 3 × 3 window is used and then the location of the hairy pixels on the skin is accurately detected according to the gray mean value and color information. The bilinear interpolation is used to modify the gray value of the hairy pixels in order to eliminate the negative effect of noise and tiny hairs on the texture. After the above pretreatment, the gray level co-occurrence matrix (GLCM) is calculated. On the basis of this, the four characteristic parameters, including the second moment, contrast, entropy and correlation, and their mean value are calculated at 45 ° intervals. The quantitative evaluation model of skin texture based on GLCM is established, which can calculate the comprehensive parameters of skin condition. Experiments show that using this method evaluates the skin condition, both based on biochemical indicators of skin evaluation methods in line, but also fully consistent with the human visual experience. This method overcomes the shortcomings of the biochemical evaluation method of skin damage and long waiting time, also the subjectivity and fuzziness of the visual evaluation, which achieves the non-destructive, rapid and quantitative evaluation of skin condition. It can be used for health assessment or classification of the skin condition, also can quantitatively evaluate the subtle improvement of skin condition after using skin care products or stage beauty.     

A standardised protocol for texture feature analysis of endoscopic images in gynaecological cancer

Background  

In the development of tissue classification methods, classifiers rely on significant differences between texture features extracted from normal and abnormal regions. Yet, significant differences can arise due to variations in the image acquisition method. For endoscopic imaging of the endometrium, we propose a standardized image acquisition protocol to eliminate significant statistical differences due to variations in: (i) the distance from the tissue (panoramic vs close up), (ii) difference in viewing angles and (iii) color correction.     

A network of coupled chaotic maps for adaptive multi-scale image segmentation

In this paper, a network of coupled chaotic maps for multi-scale image segmentation is proposed. Time evolutions of chaotic maps that correspond to a pixel cluster are synchronized with one another, while this synchronized evolution is desynchronized with respect to time evolution of chaotic maps corresponding to other pixel clusters in the same image. The number of pixel clusters is previously unknown and the adaptive pixel moving technique introduced in the model makes it robust enough to classify ambiguous pixels.     

Microarray-based maps of copy-number variant regions in European and sub-Saharan populations

The genetic basis of phenotypic variation can be partially explained by the presence of copy-number variations (CNVs). Currently available methods for CNV assessment include high-density single-nucleotide polymorphism (SNP) microarrays that have become an indispensable tool in genome-wide association studies (GWAS). However, insufficient concordance rates between different CNV assessment methods call for cautious interpretation of results from CNV-based genetic association studies. Here we provide a cross-population, microarray-based map of copy-number variant regions (CNVRs) to enable reliable interpretation of CNV association findings. We used the Affymetrix Genome-Wide Human SNP Array 6.0 to scan the genomes of 1167 individuals from two ethnically distinct populations (Europe, N=717; Rwanda, N=450). Three different CNV-finding algorithms were tested and compared for sensitivity, specificity, and feasibility. Two algorithms were subsequently used to construct CNVR maps, which were also validated by processing subsamples with additional microarray platforms (Illumina 1M-Duo BeadChip, Nimblegen 385K aCGH array) and by comparing our data with publicly available information. Both algorithms detected a total of 42669 CNVs, 74% of which clustered in 385 CNVRs of a cross-population map. These CNVRs overlap with 862 annotated genes and account for approximately 3.3% of the haploid human genome.We created comprehensive cross-populational CNVR-maps. They represent an extendable framework that can leverage the detection of common CNVs and additionally assist in interpreting CNV-based association studies.     

Learning regions of interest from low level maps in virtual microscopy

Virtual microscopy can improve the workflow of modern pathology laboratories, a goal limited by the large size of the virtual slides (VS). Lately, determination of the Regions of Interest has shown to be useful in navigation and compression tasks. This work presents a novel method for establishing RoIs in VS, based on a relevance score calculated from example images selected by pathologist. The process starts by splitting the Virtual Slide (VS) into a grid of blocks, each represented by a set of low level features which aim to capture the very basic visual properties, namely, color, intensity, orientation and texture. The expert selects then two blocks i.e. A typical relevant (irrelevant) instance. Different similarity (disimilarity) maps are then constructed, using these positive (negative) examples. The obtained maps are then integrated by a normalization process that promotes maps with a similarity global maxima that largely exceeds the average local maxima. Each image region is thus entailed with an associated score, established by the number of closest positive (negative) blocks, whereby any block has also an associated score. Evaluation was carried out using 8 VS from different tissues, upon which a group of three pathologists had navigated. Precision-recall measurements were calculated at each step of any actual navigation, obtaining an average precision of 55% and a recall of about 38% when using the available set of navigations.     

Automatic cell segmentation in cyto- and histometry using dominant contour feature points.

Automatic cell segmentation has various application potentials in cytometry and histometry. In this paper, an automatic cluster (touching) cell segmentation approach using the dominant contour feature points has been presented. Dominant feature points are the locations of indentation on the contour of the cluster. First, dominant feature points on the contour of the cluster are detected by distance profile. Next, using shape features of the cells, these feature points are selected for segmentation. We compared the results of the proposed method with manual segmentation and observed that the method has an overall accuracy about to 82%.     

Motion edges and regions guide image segmentation by colour.

Image segmentation is an important early stage in visual processing in which the visual system groups together parts of the image that belong together, prior to or in conjunction with object recognition. Two principal processes may be involved in image segmentation: an edge-based process that uses feature contrasts to mark boundaries of coherent regions, and a region-based process that groups similar features over a larger scale. Earlier, we have shown that motion and colour interact strongly in image segmentation by the human visual system. Here we explore the nature of this interaction in terms of edge- and region-based processes. We measure performance on a region-based colour segmentation task in the presence of distinct types of motion information, in the form of edges and regions which in themselves do not reveal the location of the colour target. The results show that both motion edges and regions may guide the integrative process required for this colour segmentation task. Motion edges appear to act by delimiting areas over which to integrate colour information, whereas motion similarities define primitive surfaces within which colour grouping and segmentation processes are deployed.     

Self-organizing maps for visual feature representation based on natural binocular stimuli

We model the stimulus-induced development of the topography of the primary visual cortex. The analysis uses a self-organizing Kohonen model based on high-dimensional coding. It allows us to obtain an arbitrary number of feature maps by defining different operators. Using natural binocular stimuli, we concentrate on discussing the orientation, ocular dominance, and disparity maps. We obtain orientation and ocular dominance maps that agree with essential aspects of biological findings. In contrast to orientation and ocular dominance, not much is known about the cortical representation of disparity. As a result of numerical simulations, we predict substructures of orientation and ocular dominance maps that correspond to disparity maps. In regions of constant orientation, we find a wide range of horizontal disparities to be represented. This points to geometrical relations between orientation, ocular dominance, and disparity maps that might be tested in experiments. Received: 9 July 1998 / Accepted in revised form: 2 June 1999     

Semiautomated image segmentation of bone marrow biopsies by texture features and mathematical morphology

OBJECTIVE: To present the preliminary results of a method for semiautomated detection of fat and hematopoietic cells as well as trabecular surfaces in bone marrow biopsies in order to calculate the percentage of each type of tissue or cell area in relation to the whole area. STUDY DESIGN: The results were derived from selected clinical cases. Twenty-six biopsies were used, presenting varied distributions of cellularity and trabecular topography. The approach is based on digital image processing techniques and pattern recognition methods using textural features obtained from biopsy images. The results were improved with mathematical morphology filters. RESULTS: A low computational cost algorithm is obtained that produces highly satisfactory results. The method is faster and more reproducible than conventional ones, such as region growing, edge detection, splitting and merging. CONCLUSION: The results with this computer-assisted technique were compared to those obtained by visual inspection by 2 expert pathologists, and differences of < 9% were observed.     

Formation of segregated regions of feature detecting cells in self-organizing nerve field

This paper investigates the formation of a segregated region for a specific set of input signals in a nerve field. The segregated region is formed by a feature detecting cell group which fires only for a specific set of input signals. This firing region is called the region of feature detecting cells for the corresponding input set. First, a basic self-organizing model is given. The model is composed of the first input layer, the second nerve cell layer with lateral inhibitory interactions (it is called a lateral-inhibition type nerve field) and an inhibitory nerve cell. An input signal is an-dimensional vector whose components assume continuous values. Next, the condition which gurantees the formation of a segregated region for a specific set of input signals is derived, and the properties of the model are discussed based on the derived condition. In addition, the behavior of the model is examined through computer-simulated experiments. The following observations are made: When a certain condition is satisfied, a segregated region is formed in the nerve field according to a specific set of input signals. By varying the parameters of learning, the region is formed depending on the similarity between input signals. The regions for similar input signals are formed near each other in the nerve field. The region is created depending on the occurence probability of the input signal and its norm.