Felipe Fernández

Improving image segmentation quality through effective region merging using a hierarchical social metaheuristic

This paper proposes a new evolutionary region merging method in order to efficiently improve segmentation quality results. Our approach starts from an oversegmented image, which is obtained by applying a standard morphological watershed transformation on the original image. Next, each resulting region is represented by its centroid. The oversegmented image is described by a simplified undirected weighted graph, where each node represents one region and weighted edges measure the dissimilarity between pairs of regions (adjacent and non-adjacent) according to their intensities, spatial locations and original sizes. Finally, the resulting graph is iteratively partitioned in a hierarchical fashion into two subgraphs, corresponding to the two most significant components of the actual image, until a termination condition is met. This graph-partitioning task is solved by a variant of the min-cut problem (normalized cut) using a hierarchical social (HS) metaheuristic. We have efficiently applied the proposed approach to brightness segmentation on different standard test images, with good visual and objective segmentation quality results.

Particle filter on GPUs for real-time tracking

Efficient object tracking is required by many Computer Vision application areas like surveillance or robotics. It deals with statespace variables estimation of interesting features in image sequences and their future prediction. Probabilistic algorithms has been widely applied to tracking. These methods take advantage of knowledge about previous states of the system reducing the computational cost of an exhaustive search over the whole image. In this framework, posterior probability density function (pdf) of the state is estimated in two stages: prediction and update. General particle filters are based on discrete representations of probability densities and can be applied to any state-space model [Arulampalam et al. 2002]. Discrete particles j of a set (Xt ,Πt) = {(x0 t ,π0 t )...(xN t ,πN t )} in time step t, contains information about one possible state of the system x j t and its importance weight π j t . In a practical approach, particle weights computation is the most expensive stage of the particle filter algorithm, and it has to be executed at each time step for every particle [Deutscher et al. 2000].

(Bi, R)2O3 (R : Nd, Sm and Dy) oxides as potential pigments

In this work we have tackled the feasibility of the colored oxides in the Bi 2 O 3 -R 2 O 3 system as ecological inorganic pigments. We have prepared by solid state reaction the mixed oxides (Bi 2 O 3 ) 1-x (R 2 O 3 ) x , R: Nd, Sm and Dy, with nominal compositions: Nd, x=0.2 and 0.5; Sm, x=0.4; and Dy, x=0.35. The obtained orange (Bi 2 O 3 ) 0.6 (Sm 2 O 3 ) 0 4 and (Bi 2 O 3 ) 0.65 (Dy 2 O 3 ) 0.35 oxides present f.c.c. fluorite-type structure. (Bi 2 O 3 ) 0.8 (Nd 2 O 3 ) 0.2 oxide consists of a mixture of two phases which present f.c.c. δ-Bi 2 O 3 and Bi-Sr-O-type structures in the 900-1200°C temperature range. In the synthesis of the oxide with nominal composition (Bi 2 O 3 ) 0.5 (Nd 2 O 3 ) 0.5 we have obtained a greenish sample with higher neodymium content. This sample presents the Bi 3 R 5 O 12 -type structure. Diffuse reflectance data and color coordinates suggest that the (Bi 2 O 3 ) 0.6 (Sm 2 O 3 ) 0.4 and (Bi 2 O 3 ) 0.65 (Dy 2 O 3 ) 0.35 oxides are expected to be promising candidates as new ecological pigments.

A low-level hybridization between memetic algorithm and VNS for the max-cut problem

The Max-Cut problem consists of finding a partition of the graph nodes into two subsets, such that the sum of the edge weights having endpoints in different subsets is maximized. This NP-hard problem for non planar graphs has different applications in areas such as VLSI and ASIC design. This paper proposes an evolutionary hybrid algorithm based on low-level hybridization between Memetic Algorithms and Variable Neighborhood Search. This algorithm is tested and compared with the results, found in the bibliography, obtained by other hybrid metaheuristics for the same problem. Achieved experimental results show the suitability of the approach, and that the proposed hybrid evolutionary algorithm finds near-optimal solutions. Moreover, on a set of standard test problems, new best known solutions were produced for several instances.

Phase transitions and magnetic behaviour of R1−xCa1+xCrO4 oxides (RY or Sm) (0⩽x⩽0.5)

Abstract The RCaCrO4 (RSm or Y) stoichiometric oxides crystallize with orthorhombic symmetry and space group Bmab. In the case of the SmCaCrO4 at 1073 K a structural phase transition of the type Bmab (orthorhombic) → I4/mmm (tetragonal) has been observed. Below room temperature at about 190 K a second Bmab→P42/ncm (tetragonal) phase transition has been detected from X-ray diffraction data. The magnetic properties of both SmCaCrO4 and YCaCrO4 reveal the existence of antiferromagnetic ordering above room temperature. The appearance of the ferromagnetic component at the temperature at which the low temperature phase transition takes place has been explained as the result of the canting of the antiferromagnetically ordered chromium sublattice. A field-induced magnetic transition has been observed at 50 K and 14 kOe for SmCaCrO4. The doping of the samarium sublattice, which increases the Ca Sm ratio from 1 to 1.5 to give mixed-valence compounds of the type Sm1−xCa1+xCrO4, yields very important changes in the structural properties.

A Takagi-Sugeno model with fuzzy inputs viewed from multidimensional interval analysis

Takagi-Sugeno (T-S) fuzzy systems have been successfully applied to a wide range of problems and have demonstrated significant advantages in nonlinear control. This paper presents a fuzzified T-S interpolation-approximation system mainly based on the specification of a multidimensional crisp partition that defines the corresponding local regions (multidimensional intervals) where the corresponding rules apply, the related output functions and a reduced set of global fuzzy parameters. These fuzzy parameters capture the different uncertainties of a fuzzy system: imprecision of inputs, vagueness of antecedent linguistic labels and smoothness requirements of outputs. This approach makes easier the design of a zero-order product-sum T-S system with fuzzified inputs, fuzzified antecedent crisp partition, and outputs with an additional spatial output filter. Convolution operations applied on an equalized and normalized input domain are considered to specify the corresponding fuzzification of a crisp partition. The kernels of these convolutions are even B-spline functions of order n, constructed from a n-fold convolution of an even interval characteristic function. We use a correctness-preserving transformation to simplify the output computation: a global transformation of imprecision of inputs, vagueness of antecedent terms and smoothness requirements of outputs into a set of off-line convolution operations applied to the corresponding antecedent crisp partition. By this method a fuzzified zero-order T-S system defined on a multidimensional crisp partition is directly transformed into a multidimensional spline interpolator-approximator by means of fuzzy operations and an equalization-normalization of the corresponding input domain.

Improving GPU particle filter with shader model 3.0 for visual tracking

Human-Computer Interaction is evolving towards non-contact devices using perceptual user interfaces. Recent research in human motion analysis and visual object tracking make use of the Particle Filter (PF) framework. The PF algorithm enables the modeling of a stochastic process with an arbitrary probability density function, by approximating it numerically with a set of samples called particles. The DirectX Shader Model is a common framework for accessing graphics hardware features in terms of shading functionality. In particular, Shader Model 3.0 compliant graphics cards must support features such as dynamic branching, longer shader programs and texture lookups from vertex buffers, among others. In this work, we propose new improvements on previous CPU/GPU Particle Filter frameworks [Montemayor et al. 2004; Lanvin et al. 2005]. In particular, we have reduced bandwidth requirements in the data allocation stage using GPU texture reads instead of CPUGPU memory transfers. But more importantly, using new features in Shader Model 3.0 we can move all the previous particle filtering CPU stages to the GPU, keeping all the computation on the video card and avoiding expensive data readback.

Top-Down Evolutionary Image Segmentation Using a Hierarchical Social Metaheuristic

This paper presents an application of a hierarchical social (HS) metaheuristic to region-based segmentation. The original image is modelled as a simplified image graph, which is successively partitioned into two regions, corresponding to the most significant components of the actual image, until a termination condition is met. The graph-partitioning task is solved as a variant of the min-cut problem (normalized cut) using an HS metaheuristic. The computational efficiency of the proposed algorithm for the normalized cut computation improves the performance of a standard genetic algorithm. We applied the HS approach to brightness segmentation on various synthetic and real images, with stimulating trade-off results between execution time and segmentation quality.

Application of multidimensional retiming and matroid theory to DSP algorithm parallelization

This paper presents a novel optimization methodology to implement parallel DSP algorithms based on the application of general multidimensional retiming techniques and graphic matroid theory. These methods improve the throughput of a synchronous circuit and guarantee that all functional elements work in parallel. This approach is also used to achieve full parallelism in multiprocessor architecture. Three multidimensional retiming methods: node, cocycle and cycle are introduced to assist the digital designer in DSP algorithms parallelization. The main theoretical advantage of this approach is a deeper formal treatment of the set of data dependences for the considered algorithms. This orientation also provides an intuitive tool for the graphical analysis and design of the corresponding systems, and a mathematical tool for analysing and transforming (retiming) the involved algorithms. Application of these matrix optimization techniques to the optimal delay management problem for multidimensional DSP algorithms is also shown.

Region merging for severe oversegmented images using a hierarchical social metaheuristic

This paper proposes a new evolutionary region merging method to improve segmentation quality result on oversegmented images. The initial segmented image is described by a modified Region Adjacency Graph model. In a second phase, this graph is successively partitioned in a hierarchical fashion into two subgraphs, corresponding to the two most significant components of the actual image, until a termination condition is met. This graph-partitioning task is solved as a variant of the min-cut problem (normalized cut) using a Hierarchical Social (HS) metaheuristic. We applied the proposed approach on different standard test images, with high-quality visual and objective segmentation results.