Multispectral image segmentation based on Cartesian complexes and their associated oriented matroids

Presents a new low-complexity technique for the segmentation of multispectral images, based on the use of a tree-structured Markov random field model. The image is associated with a binary tree, and is segmented recursively through a sequence of local splits based on a maximum a posteriori probability rule. To improve the reliability of the process, merging of nodes is now considered besides splitting, so as to allow for the re-shaping of incorrect region boundaries. Experimental results show that the new algorithm increases the fitness of the segmentation to the actual features of the image.

Low-level segmentation of multispectral images via agglomerative clustering of uniform neighbourhoods

We present a parallel mean shift algorithm running on CUDA and its possible application in segmentation of multispectral images. The aim of this paper is to present a method of analyzing highly noised multispectral images of various objects, so that important features are enhanced and easier to identify. The algorithm finds applications in analysis of multispectral images of eyes so that certain features visible only in specific wavelengths are made clearly visible despite high level of noise, for which processing time is very long.

Diseases in moringa leaves pose a serious threat to PT. Marada Kelor Sumba, a moringa processing company located in Kelurahan Temu, Kanatang District, East Sumba Regency. The main products from moringa processing are moringa powder and moringa tea bags, which are highly beneficial as supplementary food for infants and pregnant women. The company collaborates with the East Sumba Government for the prevention and handling of stunting. The primary objective of this study is to optimize image segmentation methods to identify diseases in moringa leaves with significant implications using MATLAB software. The image segmentation method using Thresholding is applied to separate diseased moringa leaves from healthy ones. Additionally, filters are applied to enhance disease image segmentation capabilities and manipulate images. This research focuses on the image segmentation stage using MATLAB. The benefits of this research are crucial in agriculture and technology development, including early segmentation of diseases in moringa leaves, improvement of disease image segmentation process efficiency, contribution to image-based agricultural technology, and enrichment of scientific knowledge and understanding of disease image segmentation in moringa plants. The segmentation accuracy obtained from testing all samples is 66,67%.

This paper describes a printed Thai character recognition method which uses topological properties. The recognition system is divided into two stages. In the rough classification stage, a thinning technique is applied to eliminate noise and fine features. The thinned characters are separated into cluster domains by using a linear decision function based on the height and width of each character. In the fine classification stage, the ‘connected numbers’ of points in the character pattern are computed. Topological properties are then used to distinguish between the characters.

This paper presents a novel method for segmentation and change detection of multispectral images using proximal splitting-based clustering and multiclass support vector machine (MSVM). Initially, the multitemporal satellite images are preprocessed and then textures are extracted using Difference of Offset Gaussian filter. In general, the traditional clustering method uses Euclidean distance as a prime factor for segmentation process. For multitextured images such as remotely sensed images, this metric provides inconsistent output. To achieve better segmented results, proximal splitting algorithm has been proposed. This method has been considered as a solution for iterative minimization problem, which is required to find exact changes between the multitemporal images. The MSVM is chosen to group the segmented clusters into a fixed number of classes, since the clusters obtained from the proximal splitting algorithm are not independent with each other. Then, the classified images are subjected to image differencing method to detect the changes. Experimentation is performed with two real data sets of Landsat7 images, which illustrates that the mean of difference in area obtained by the proposed method is reduced by an average of 35.24% compared to the conventional system. The validity index obtained for data set 1 using proposed algorithm is lower than the existing methods.

In this paper; we introduce a system of automatic recognition of characters based on the Random Forest Method in non-constrictive pictures that are stemmed from the terminals Mobile phone. After doing some pretreatments on the picture, the text is segmented into lines and then into characters. In the stage of characteristics extraction, we are representing the input data into the vector of primitives of the zoning types, of diagonal, horizontal and of the Zernike moment. These characteristics are linked to pixels' densities and they are extracted on binary pictures. In the classification stage, we examine four classification methods with two different classifiers types namely the multi-layer perceptron (MLP) and the Random Forest method. After some checking tests, the system of learning and recognition which is based on the Random Forest has shown a good performance on a basis of 100 models of pictures.

In this paper we employ the Integrated Active Contour (IAC) model of vector-value image segmentation and C-V model based on PDE in the segmentation of two channels DMSP nighttime satellite cloud images. The results of segmentation show that the IAC model can integrate the information of each channel throughout the segmentation and can be employed in the image with uneven background and output a uniform segmentation result. The IAC model is fitted for multi-spectral satellite image segmentation without much prior information.

The design, implementation, and preliminary testing of a computer system for automatic multispectral magnetic resonance imaging analysis is presented. The modular structure of the system permits easy comparison between various classification algorithms. The classification accuracy of traditional statistical pattern-recognition algorithms is compared to the results that can be obtained with neural networks of different topologies. Quantitative (confusion matrices) as well as visual (segmented images) results of a study performed on sets of normal and pathological images are presented. Images segmented with a neural network classifier (NNC) appear less noisy than images segmented with a maximum likelihood classifier (MLC), and it has been observed that the NNC is less sensitive to the selection of the training sets than the MLC

In recent years, models of spatial relations, especially topological relations, have attracted much attention from the GIS community. In this paper, some basic topologic models for spatial entities in both vector and raster spaces are discussed. It has been suggested that, in vector space, an open set in 1-D space may not be an open set any more in 2-D and 3-D spaces. Similarly, an open set in 2-D vector space may also not be an open set any more in 3-D vector spaces. As a result, fundamental topological concepts such as boundary and interior are not valid any more when a lower dimensional spatial entity is embedded in higher dimensional space. For example, in 2-D, a line has no interior and the line itself (not its two end-points) forms a boundary. Failure to recognize this fundamental topological property will lead to topological paradox. It has also been stated that the topological models for raster entities are different in Z^{2} and R^{2}. There are different types of possible boundaries depending on the definition of adjacency or connectedness. If connectedness is not carefully defined, topological paradox may also occur. In raster space, the basic topological concept in vector space—connectedness—is implicitly inherited. This is why the topological properties of spatial entities can also be studied in raster space. Study of entities in raster (discrete) space could be a more efficient method than in vector space, as the expression of spatial entities in discrete space is more explicit than that in connected space.

The nonlinear dimensionality reduction and its effects on vector classification and segmentation of hyperspectral images are investigated in this letter. In particular, the way dimensionality reduction influences and helps classification and segmentation is studied. The proposed framework takes into account the nonlinear nature of high-dimensional hyperspectral images and projects onto a lower dimensional space via a novel spatially coherent locally linear embedding technique. The spatial coherence is introduced by comparing pixels based on their local surrounding structure in the image domain and not just on their individual values as classically done. This spatial coherence in the image domain across the multiple bands defines the high-dimensional local neighborhoods used for the dimensionality reduction. This spatial coherence concept is also extended to the segmentation and classification stages that follow the dimensionality reduction, introducing a modified vector angle distance. We present the underlying concepts of the proposed framework and experimental results showing the significant classification improvements

Most cartographic work is made extracting features from aerial or space images. A first step in this work is segmenting the images in regions that represent, as close as possible, cartographic entities (e.g., roads, buildings, vegetation). Region-Growing segmentation is implemented in a multispectral image using an open source programming language. This segmentation method is analyzed for land used and land cover applications, and it is compared with classification-based segmentation, known as Fuzzy K-Means. Both algorithms, Region Growing and Fuzzy K-Means, are run in an aerial image with four spectral bands (red, green, blue, and near infrared). Depending on the scale, the values of the parameters of the algorithms can yield an under segmentation or over segmentation results. Advantages and disadvantages of both segmentation methods are provided.

ABSTRACTThis paper presents a multi-objective fuzzy clustering approach under iterative modelling for the segmentation of human brain multispectral magnetic resonance image. It optimizes the global spatial fuzzy compactness and fuzzy spatial separation among clusters. Our method generates an ensemble of Pareto solutions among which the final solution will be proposed to the analyst. At this final step, the Shapley value concept is utilized to select the best solution. An empirical study on brain images shows a better performance of our proposed method when compared to competing algorithms.

High-resolution multi-spectral remote sensing image of urban areas provides both structural and spectral information about urban scenes. In segmentation of such complex image scenes, very thin, env ...

In this work a technique for cloud classification from MSG-SEVIRI (Meteosat Second Generation-Spinning Enhanced Visible and Infra-red Imager) imagery is presented. It is based on the segmentation of the multispectral images using local texture information. The goal is to study, how local texture measures, can distinguish different cloud genres, and to develop a system based on multiresolution analysis that combines spectral and textural image characteristics in a multiscale model, to facilitate the automated interpretation of satellite cloud images. To evaluate the performance of this analysis we applied the methodology to a brodatz database and to MSG images. A series of 168 infrared and visible satellite images from the geostationary satellite METEOSAT were employed. Three different texture feature sets (a total of 24 features) were extracted using the cooccurrence matrices in the wavelet domain, and the fuzzy c-means clustering method is used for the classification. The approach indicated very good performance in classification experiment. The proposed feature extraction and classification scheme achieved high accuracies, with an overall classification accuracy of 97.53%.