Image retrieval by color regions

With the widespread availability of personal computers and the staggering amount of digital imagery available, recent investigation has focused on the storage, query and retrieval of images from large image databases. Methods, which are employed presently, precompute image indices, which are deemed to be statistical representations of image information. Queries of colour, shape, texture, etc., are then performed directly on these indices to find valid images. The authors propose a new indexing technique which calculates the multidimensional histogram of the directional detail in a given image. They apply wavelet theory and multiresolution analysis to extract the directional information from an image and then map this information into 3-dimensional vectors. Histograms of these vectors are then calculated at different levels of resolution, allowing progressive image query using colour histogram techniques to be directly applied. The technique is capable of querying and searching a database for images based on attributes such as smoothness, randomness and horizontal, vertical and diagonal edges at varying resolutions.

Retrieval by contents in image databases is a subject of growing relevance in many application areas. In this task, visual techniques permit the introduction of natural criteria of similarity in the specification of queries. Measures of similarity address the visual impressions that images provide to the user, the possibly incomplete knowledge about image contents and the cognitive process by which picture contents are interpreted. In this paper, iconic retrieval of images is presented based on an original language for the specification of spatial relationships between the imaged objects, referred to as Spatial Logic. With Spatial Logic the subjective user's point of view in expressing the query is supported in two distinct regards: the perceptual and cognitive processes of the user about spatial relationships between the objects are taken into account by considering 3D symbolic descriptions of imaged scenes; the incomplete knowledge by the user is considered by permitting different levels of detail in the description of the spatial relationships. A system which exploits Spatial Logic to support image retrieval by contents through iconic interaction is presented.

The problem of simulating light scattering patterns of polymerization induced phase separation processes by the mechanism of nucleation and growth is addressed with the objective of reducing the number of computations required to carry out the simulation. A method, based on the Fast Fourier Transform (FFT), that takes advantage of the spherical nature of the phase separated domains, is proposed. Contrary to the case in which the spherical particles are randomly placed in arbitrary locations in the scattering volume, the proposed computational scheme requires the locations not to be arbitrary but to belong to a previously defined equally spaced grid. This last constraint determines levels of resolution in which the simulation can be performed. The FFT based method is tested with a morphology previously generated with a model of the phase separation process. The light scattering pattern corresponding to this morphology, mainly represented through the particle size distribution of the spherical separated domains, has been simulated at different levels of resolution of the grid placed in the scattering domain and of the sizes of the domains. It has been found that the original morphology represented by a distribution with 17 different sizes can be approximated by a smaller number, while keeping the total volume of particles the same, without a noticeable loss of precision. This, together with the fact that the computations can be performed at different levels of resolution, gives the chance to explore an important number of approximations to the problem, many of which produce, using the proposed method, results that are comparable to those obtained at higher levels of resolution, with the advantage of requiring a smaller number of computations.

The properties of water at physiological conditions can be modelled at different levels of resolution: (1) sub-atomic models that take into account electronic degrees of freedom, (2) atomic models that only account for atomic degrees of freedom and (3) supra-molecular models that only involve some supra-molecular degrees of freedom. To enhance the computational efficiency of molecular simulation, models at different levels of resolution should be simultaneously usable for different parts of a system for which the level of detail of interest is different. This requires these different types of models to be compatible with each other. In the present study, the compatibility of two polarisable models for liquid water, COS/G2 and COS/D, with a non-polarisable model for liquid water, SPC, is investigated. It is shown that these models are compatible. The polarisable models can thus be used to solvate biomolecules described by a biomolecular force field that is compatible with the SPC water model.

The goal of this paper is to present a quantitative study about Telco churn analysis and the inherent problems of its application and computer processing. An independent customer characterizes the approach and the pre-processing step uses the wavelet transform concept; the pattern classification problem is solved by an artificial neural network (RBFRadial Basis Function). The signal representation considers different levels of resolution (multi-resolution). This idea is frequently used in image/video processing where general image versions with a low resolution are stored in different levels of resolution. The concepts of multi-resolution have been implemented by a digital filter bank, Hi-pass and Low-pass. Each sub-band will give the best resolution, for each spectral region. The idea is to look for a spectral transient. A scalable Daubechies’ function db 2 and db 3 were tested for this study and the results were used as input for the RBF neural network. Daubechies’ function has shown important properties as for instance, orthogonality and scalability, which are not presented by Morlet’s function. Some results of the experiment were presented to point out the performance of the wavelet pre-processing approach.

Pore-scale modeling of flow in porous and fractured media is a first key step to provide a firm knowledge basis for effective aquifer management practices in environmental applications or to improved reservoir performance in the oil and gas field. We present a performance comparison of two numerical approaches for the pore-scale simulation of fully saturated flow in natural porous media. Real samples of rocks at the millimeter scale whose complex pore space geometry is reconstructed through X-ray computer tomography are considered. The governing flow equation in the system are solved by employing a Body-Fitted (BF) grid and an Immersed Boundary (IB) method, as respectively embedded in FLUENT and EULAG (Prusa et al., 2008) software environments. The BF technique allows solving the flow problem within the real pore space and is widely adopted in computational fluid dynamics since it allows to fit complex geometries while preserving accuracy, up to a certain level of detail. The method is associated with some limitations in the presence of extremely complex geometrical setting of the kind encountered in natural or reconstructed micro scale porous systems, mainly due to the time-demanding mesh generation stage. On the other hand, IB methods can take full advantage of uniform Cartesian grid implementations describing both the fluid and solid regions in the domain. Grain size distributions are modeled by considering an additional forcing term in the governing flow equation. This term inhibits Sow through the solid matrix and hence acts to enforce the no-slip condition at solid walls. The effectiveness of such term, i.e., its ability to prevent mass flow rate within the solid domain, is controlled by a parameter, a, which is directly related to the characteristic time scale of the forcing action. First, the performance of each model is analyzed separately on the same rock samples. Grid dependence is tested for the BF method upon comparing simulations at three different levels of resolution. A detailed sensitivity analysis is performed to assess the role of the parameter α in the IB method simulations. Finally, the two models are compared in terms of resulting flow fields, mass-conservation capability, hydraulic conductivity estimates and computational resources requirement.

Defining a unifying model for describing spatial data at different levels of resolution is a relevant issue in several applications involving spatial data handling. Also, dimension-independence is becoming fundamental in many application contexts. We propose a unifying model for multiresolution description of spatial data which works in arbitrary dimension. Graph-based representations for encoding different instances of the abstract model are described.

Avian communities living in a complex environment were studied by censuses of singing males, using several levels of resolution which focussed on different properties of the community. Within each level of resolution, various community parameters in two grassland habitat types were calculated over the entire breeding season, and for census intervals within the breeding season. The community level of resolution focussed on community—wide patterns, such as total density and total biomass, in order to reveal constraints operating on all species in each habitat. An intermediate level of resolution focussed on relative properties of species such as relative abundances, species richness, and dominance concentration. At the individual species level of resolution, densities and identities of species were the properties upon which the analysis focussed. These techniques allowed characterization of responses of the avian communities to complex changes in the system in which they were embedded. Long—term averages in monthly rainfall on the sites used for this study indicated a bimodal distribution of rainfall. The first peak occurred in December and the second in July and August. During the two years of study, summer rainfall was similar, but winter rainfall in 1983 was °2—3 times as great as in 1982. Avian communities in habitats with mesquite trees (mesquite savannah) responded differently than communities in grassland habitats, and these responses could be seen at different levels of resolution. Peak densities and biomass occurred during May—June in mesquite savannah habitats, while grassland communities achieved highest densities and biomass during July and August. Average density per individual showed little seasonal pattern, but was lower for mesquite savannah communities than for grassland communities. Species richness increased in mesquite savannah habitats in 1983, but decreased in grassland habitats that year. Patterns of evenness and dominance concentration suggested that in 1983, rare species of large body size were added to the community in mesquite savannah, while in grassland habitats, species already present during 1982 increased in density. A community ordination revealed that mesquite savannah habitats were dominated numerically by small—bodied insectivorous species during the early part of the breeding season but became similar to grassland communities in July and August. Grassland communities were dominated by large—bodied emberizine finches. These characteristics of the avian communities corresponded closely to the phenology of important primary producers. Mesquite produces leaves high in nutritional value and produces flowers during April—June; thereafter older leaves decline in nutrient content and flowers disappear. These changes in mesquite productivity undoubtedly produce corresponding changes in insect communities exploitable by small—bodied insectivores. Perennial grasses concentrate aboveground productivity to July—August and insect communities associated with grasses increased in biomass during this period, thus increasing resources available for breeding. These results suggest that avian community responses to complex changes in their environments can be identified and related to outside inputs into the environment, such as rainfall. Identification of community responses is facilitated by emphasis on several observational scales. Communities appear individualistic to observers because (1) adaptational units of species are probably much larger than study areas; (2) species ability to adapt to local conditions might be diluted by gene flow among populations; and (3) if most evolutionary change occurs during speciation, then species adaptations will reflect a relatively narrow range of selection pressures that operated briefly on the species during the past.

One of the main challenges in medical image retrieval is the increasing volume of image data, which render it difficult for domain experts to find relevant information from large data sets. Effective and efficient medical image retrieval systems are required to better manage medical image information. Text‐based image retrieval (TBIR) was very successful in retrieving images with textual descriptions. Several TBIR approaches rely on models based on bag‐of‐words approaches, in which the image retrieval problem turns into one of standard text‐based information retrieval; where the meanings and values of specific medical entities in the text and metadata are ignored in the image representation and retrieval process. However, we believe that TBIR should extract specific medical entities and terms and then exploit these elements to achieve better image retrieval results. Therefore, we propose a novel reranking method based on medical‐image‐dependent features. These features are manually selected by a medical expert from imaging modalities and medical terminology. First, we represent queries and images using only medical‐image‐dependent features such as image modality and image scale. Second, we exploit the defined features in a new reranking method for medical image retrieval. Our motivation is the large influence of image modality in medical image retrieval and its impact on image‐relevance scores. To evaluate our approach, we performed a series of experiments on the medical ImageCLEF data sets from 2009 to 2013. The BM25 model, a language model, and an image‐relevance feedback model are used as baselines to evaluate our approach. The experimental results show that compared to the BM25 model, the proposed model significantly enhances image retrieval performance. We also compared our approach with other state‐of‐the‐art approaches and show that our approach performs comparably to those of the top three runs in the official ImageCLEF competition.

Characterization of shaking intensity distribution and seismic assessment of RC buildings for the Kashmir (Pakistan) earthquake of October 2005

The conversion of coarse-grained to atomistic models is an important step in obtaining insight about atomistic scale processes from coarse-grained simulations. For this process, called backmapping or reverse transformation, several tools are available, but these commonly require libraries of molecule fragments or they are linked to a specific software package. In addition, the methods are usually restricted to specific molecules and to a specific force field. Here, we present an alternative method, consisting of geometric projection and subsequent force-field based relaxation. This method is designed to be simple and flexible, and offers a generic solution for resolution transformation. For simple systems, the conversion only requires a list of particle correspondences on the two levels of resolution. For special cases, such as nondefault protonation states of amino acids and virtual sites, a target particle list can be specified. The mapping uses simple building blocks, which list the particles on the different levels of resolution. For conversion to higher resolution, the initial model is relaxed with several short cycles of energy minimization and position-restrained MD. The reconstruction of an atomistic backbone from a coarse-grained model is done using a new dedicated algorithm. The method is generic and can be used to map between any two particle based representations, provided that a mapping can be written. The focus of this work is on the coarse-grained MARTINI force field, for which mapping definitions are written to allow conversion to and from the higher-resolution force fields GROMOS, CHARMM, and AMBER, and to and from a simplified three-bead lipid model. Together, these offer the possibility to simulate mesoscopic membrane structures, to be transformed to MARTINI and subsequently to an atomistic model for investigation of detailed interactions. The method was tested on a set of systems ranging from a simple, single-component bilayer to a large protein-membrane-solvent complex. The results demonstrate the efficiency and the efficacy of the new approach.

Multi-level coarse-grain model of the DEM

‘A visual idea is more powerful than verbal idea’, ‘A picture is worth more than ten thousand words’, ‘No words can convey what a picture speaks’, ‘A picture has to be seen and searched as a picture only’ are few of the well-known sayings that imply the certainty for the widespread availability of images. Common sense evidence suggests that images are required for a variety of reasons, like, illustration of text articles, conveying information or emotions that are difficult to describe in words, display of detailed data for analysis (medical images), formal recording of design data for later use (architectural plans) etc. The advent of digital photography combined with decreasing storage and processing cost, allows more and more people to have their personal collection of photographs and other visual content available on the internet. Organising these digital images into a small number of categories and providing effective indexing is imperative for accessing, browsing and retrieving useful data in “real time”. The process of digitization does not in itself make image collections easier to manage. Some form of indexing (cataloguing) is still necessary. People’s interest to have their own digital libraries has burgeoned and hence requires a data structure to preserve the images for a long time and also provide easy access to the desired images. These requirements have indeed forced the design of specialized imaging systems/ image databases, such that an access to any image is effective and efficient. An efficient image archival and retrieval system is characterized by its ability to retrieve relevant images based on their visual and semantic contents rather than using simple attributes or keywords assigned to them. Thus, it is necessary to support queries based on image semantics rather than mere-pixel-to-pixel matching. An image archival and retrieval system should therefore allow adequate abstraction mechanisms for capturing higher level semantics of images in order to support content addressability as far as possible. That is, for two images to be similar, not only the shape, color and texture properties of individual image regions must be similar, but also they must have the same arrangement (i.e., spatial relationships) in both the images. In fact, this is the strategy, which is generally being employed by our vision system most of the times. An effective method of representing images depends on the perception of knowledge embedded in images in terms of objects/components (generally known as elements) present in them along with their topological relationships. The perception of topological relationships, especially spatial relationships existing among the significant elements of an image, helps in making the image database system more intelligent, fast and flexible. An obvious method to search an image database is sequential scanning. The query is matched with all stored images (i.e., the representation of the query is matched with all representations stored in the image database) one by one. Retrievals may become extremely slow, especially when database search involves time consuming image matching operations. To deal with slow retrieval response times, and high complexity matching, an image database must utilize indexing methods that are faster than sequential scanning methods. In traditional image database systems, the use of indexing to allow database accessing has been well established. Analogously, image indexing techniques have been studied during the last decade to support representation of pictorial information in an image database and also to retrieve information from an image database. The use of significant elements present in images along with their topological relationships as indexes is the basic issue of the indexing methodologies developed to this aim.

The proliferation in volume of digital image data has exacerbated the general image retrieval problem, creating a need for efficient storage and flexible retrieval of vast amounts of image data (Chang, 1989). Whilst there have been significant technological advances with image data capture and storage, developments in effective image retrieval have not kept pace. Research in image retrieval has been divided into two areas: concept-based image retrieval and content-based image retrieval. The former focuses on the use of classification schemes or indexing terms to retrieve images while the latter focuses on the visual features of the image, such as colour, shape, texture, and spatial relationships.

This paper describes a document image analysis system using multiple agents working on a pyramid structure to separate text from graphics in the image. Text strings appear as different groupings of connected components at different image resolutions. As such, the pyramid structure, which is a multi-resolution image representation, provides a natural means of identifying and grouping of character strings in the document at different levels of resolution. The pyramid structure is also amenable to parallel processing, where multiple agents in the system can individually and concurrently look for groups of connected components at appropriate levels. The agent-based pyramid operations do not require expensive feature analysis among different connected components to detect text strings as found in other existing works.