Search

Altres ajuts: CERCA Programme/Generalitat de Catalunya Color description is one of the fundamental problems of image understanding. One of the popular ways to represent colors is by means of color names. Most existing work on color names focuses on only the eleven basic color terms of the English language. This could be limiting the discriminative power of these representations, and representations based on more color names are expected to perform better. However, there exists no clear strategy to choose additional color names. We collect a dataset of 28 additional color names. To ensure that the resulting color representation has high discriminative power we propose a method to order the additional color names according to their complementary nature with the basic color names. This allows us to compute color name representations with high discriminative power of arbitrary length. In the experiments we show that these new color name descriptors outperform the existing color name descriptor on the taskof visual tracking, person re-identification and image classification.

In order to improve the performance of the integrated navigation system, a new navigation resource management method based on an efficiency function is proposed in this paper. Firstly, the framework of the navigation resource management method is established. Secondly, the representation of the efficiency function is given, which based on the navigation accuracy, the sensor accuracy and the navigation task. Finally, this method is simulated and compared to the traditional data fusion method. The simulation results show that the positioning accuracy and the computing time of this method are improved by 20.1% and 26.7%, respectively. This method makes full use of the sensor accuracy, and manages the navigation resource effectively. It effectively improves the navigation accuracy and the computational efficiency.

We present a novel method to select waypoints from aerial images of candidate flying regions via matching suitability analysis, which is based on visual attention mechanism and feature attribute classification. At first, visual attention mechanism is used to get the saliency map of the initial image by low-rank recovery and sparse coding. The salient regions are selected to be as preparatory results with threshold constraint and nonmaxima suppression. Then we use support vector machine (SVM) to divide the preparatory results into two classes for suitable or unsuitable waypoints based on their feature attributes, which can be represented by two edge-based descriptors and two correlation-based descriptors. The experimental results show that the proposed method is valid and effective.

A new, efficient dynamic kinetic resolution (DKR) of secondary aromatic alcohols using self-made solid super acid TiO 2 / SO 4 2 - as a racemization catalyst was developed. Low-cost and easily produced TiO 2 / SO 4 2 - showed excellent racemization activity. When coupled with the lipase, Novozym 435, good biocompatibility was observed, and optically pure aromatic acetate (>99%) was obtained with a high yield. It is noteworthy that the system could be reused more than 10 times with little loss of yield or reduction in the enantiomeric excess (ee) value of product.

The technology of vision-based probe-and-drogue autonomous aerial refueling is an amazing task in modern aviation for both manned and unmanned aircraft. A key issue is to determine the relative orientation and position of the drogue and the probe accurately for relative navigation system during the approach phase, which requires locating the drogue precisely. Drogue detection is a challenging task due to disorderly motion of drogue caused by both the tanker wake vortex and atmospheric turbulence. In this paper, the problem of drogue detection is considered as a problem of moving object detection. A drogue detection algorithm based on low rank and sparse decomposition with local multiple features is proposed. The global and local information of drogue is introduced into the detection model in a unified way. The experimental results on real autonomous aerial refueling videos show that the proposed drogue detection algorithm is effective.

Recent developments of next generation sequencing technologies have led to rapid accumulation of 16S rRNA sequences for microbiome profiling. One key step in data processing is to cluster short sequences into operational taxonomic units (OTUs). Although many methods have been proposed for OTU inferences, a major challenge is the balance between inference accuracy and computational efficiency, where inference accuracy is often sacrificed to accommodate the need to analyze large numbers of sequences. Inspired by the hierarchical clustering method and a modified greedy network clustering algorithm, we propose a novel multi-seeds based heuristic clustering method, named MSClust, for OTU inference. MSClust first adaptively selects multi-seeds instead of one seed for each candidate cluster, and the reads are then processed using a greedy clustering strategy. Through many numerical examples, we demonstrate that MSClust enjoys less memory usage, and better biological accuracy compared to existing heuristic clustering methods while preserving efficiency and scalability.

A new approach of smoothing the white noise for nonlinear stochastic system was proposed. Through presenting the Gaussian approximation about the white noise posterior smoothing probability density function, an optimal and unifying white noise smoothing framework was firstly derived on the basis of the existing state smoother. The proposed framework was only formal in the sense that it rarely could be directly used in practice since the model nonlinearity resulted in the intractability and infeasibility of analytically computing the smoothing gain. For this reason, a suboptimal and practical white noise smoother, which is called the unscented white noise smoother (UWNS), was further developed by applying unscented transformation to numerically approximate the smoothing gain. Simulation results show the superior performance of the proposed UWNS approach as compared to the existing extended white noise smoother (EWNS) based on the first-order linearization.

In this paper, a novel weight multi-focus image fusion method based on the non-subsampled shearlet transform (NSST) is proposed. Firstly, the NSST is employed to decompose the source image into low frequency and high frequency sub-band coefficients with different-scale and different-directional; Secondly, the local least root mean square error (RMSE) is used as a new fusion weight for the fusion of low frequency sub-band coefficients; In order to effectively preserve the edges of the image, a novel fusion rule based on edge preserving weight is proposed to fuse high frequency sub-band coefficients. Finally, the fused image can be obtained through the inverse NSST on the fused coefficients. The experimental results show that the fused image not only contains rich details, but also preserves the notable structural features. Compare with the current state-of-the-art fusion methods, it produces better visual effect and quality assessment.

New sigma point filtering algorithms, including the unscented Kalman filter (UKF) and the divided difference filter (DDF), are designed to solve the nonlinear filtering problem under the condition of correlated noises. Based on the minimum mean square error estimation theory, the nonlinear optimal predictive and correction recursive formulas under the hypothesis that the input noise is correlated with the measurement noise are derived and can be described in a unified framework. Then, UKF and DDF with correlated noises are proposed on the basis of approximation of the posterior mean and covariance in the unified framework by using unscented transformation and second order Stirling’s interpolation. The proposed UKF and DDF with correlated noises break through the limitation that input noise and measurement noise must be assumed to be uncorrelated in standard UKF and DDF. Two simulation examples show the effectiveness and feasibility of new algorithms for dealing with nonlinear filtering issue with correlated noises.

A new and efficient dynamic kinetic resolution (DKR) process of secondary aromatic alcohols by using long carbon-chain esters as acyl donors has been developed. During the process, the transesterification catalyzed by CD8604 was found to be the main reason for the decrease in enantiomeric excess (ee). Using complex acyl donors, such as 4-chlorophenyl valerate, we could effectively inhibit the resin-catalyzed transesterification, and an excellent ee value (>99%) at high yield (>99%) was achieved. The mechanism for the inhibition of resin-catalyzed transesterification is believed to be the formation of micro-micelles in the pores of CD8604. It is noteworthy that the system can be reused more than 20 times without a loss of yield or ee value.

The main problem of particle filter (PF) in nonlinear state estimation is the particle degeneracy. Resampling operation solves degeneracy to some extent, but it results in the problem of sample impoverishment. Variance reduction technique is proposed to deal with the degeneration phenomenon in this paper, which reduces the variance of the particle weights by selecting an exponential fading factor, and this factor can be chosen adaptively and iteratively in terms of the effective particle number. A theorem is presented to show that this idea is feasible, and the procedure of this new adaptive particle filtering (APF) algorithm is presented. Then, the principle of parameter choice and the limitation of APF are discussed. Finally, a numerical example illustrates that the proposed APF has a higher estimation precision than particle filter – sampling importance resampling (PF-SIR), genetic particle filter (GPF), and particle swarm optimization particle filter (PSOPF), while the computation load of APF is mild.

A new and efficient dynamic kinetic resolution (DKR) process of secondary aromatic alcohols was developed with acid resins as racemization catalysts. Acid resin CD8604 was shown to have excellent racemization activity and good biocompatibility. When employing CD8604 and complex acyl donors as racemization catalyst and acyl donor, respectively, enantiomerically pure aromatic acetate was obtained with excellent yield and ee values through the DKR process. It is noteworthy that the system could be reused more than 10 times with little loss of yield and ee value.

In this work, a novel efficient algorithm for visual object tracking in complex conditions is proposed. The main component of this work includes two parts: Bayesian decision based confident region extraction, and mean shift iteration based tracking. A unique characteristic of the proposed algorithm is that instead of tracking the entire object, the method automatically extracts the confident region of the object through fusing multiple cues in the Bayesian framework. Those cues contain object's color feature, motion character, and dynamic surrounding color information. We tested the performance of the algorithm with video sequences under difficult conditions (complex and dynamic background, fast camera motion, object maneuvering, rotations and partial occlusion) and achieved satisfied results.

Anomaly detection in hyperspectral imagery is gaining increasing interest. However, most covariance matrix-based detectors are applied directly to all the hyperspectral data bands without considering the spectral variation and their possible different contribution to detection. Besides, the limited sample number as compared with the high dimensionality of the data can lead to the imprecise estimation of the covariance matrix and even the singularity problem. In this paper, a band-subset fuzzy integral fusion (BS-FI) detection method is presented to solve these problems. The complete set of hyperspectral data bands is first partitioned into several lower dimensional band-subsets, whose detection results are obtained separately and finally merged by a fuzzy integral fusion method. We adopt a non-parametric fuzzy support function which can utilize more statistical information and avoid the model discrepancy that might be brought in by a fixed distribution model. In addition, the fuzzy density is assigned by the ratio between the target signal and noise, which is the key to the target detection problem through an adaptive eigenvalue-based approach. In the experiments on real OMIS-I hyperspectral imagery, the proposed method outperforms the RX detector both on the complete set of bands and on each band-subset, and other band-subset fusion detectors.

Polarization vision techniques have demonstrated effectiveness in a variety of application fields including computer vision. This chapter presents 3D reconstruction and image dehazing as examples to show the benefits of polarization vision techniques.

Conventional remote sensing devices always identify a target by its reflection intensity or radiation intensity. While the polarization characteristics are always treated as noise. In the recent years, with a development of spatial remote sensing technologies, researchers have noticed the polarization characteristics which are determined by surface optical features in the process of reflection, scattering, and transmission. Measuring the polarization intensity, polarization degree, polarization angle, polarization ellipticity and emissivity can provide higher accuracy than radiation measurements. Generally, the polarization reflection spectrum contains different polarization states of different objects or different object states (such as medium, structure, surface roughness, and soil water content) which are closely related to wavelength. Polarization imaging can eliminate background noise and enhance the accuracy of target detection and identification, so that it has extensive application prospects in the field of target detection, classification, recognition, ocean remote sensing, image enhancement in severe weather conditions, computer vision, surveillance, intelligent transportation, and medical diagnostics. In this chapter, the polarization Bidirectional Reflectance Distribution Function (BRDF) is taken as an example to show the polarization characteristics of both target and background.

Multi-band polarization imaging has a variety of applications including land cover classification, military target detection, 3D surface reconstruction, and glare/shadow removal. A comprehensive utilization of spatial, spectral, and polarization information is effective to detect hidden or camouflaged targets which may not be discoverable by traditional optical imaging techniques. Present multi-band polarization sensors rely on dispersion, beam split or interference to obtain spectral information, while polarization information in different direction is acquired by changing polarization angles]. Acquisition of multiple images in different spectral bands and polarization angles may take a relatively long time, a large amount of data, and limited field of view. Conventional multi-band polarization imaging techniques may not satisfy the requirements in modern warfare in terms of dynamic environments, larger field of view, and fast response time. Recent research indicates that some aquatic organisms like dragonfly nymphs, mantis shrimps, and cuttlefish hunt with their multi-band polarization vision. Such observations sparked the studies on multi-band polarization imaging techniques.

Establishing a multi-band polarization imaging system is the premise of acquiring accurate multi-band polarization information. However, the acquired polarization and spectral data can be noisy and less accurate in the sensing process due to the factors such as dark current, positioning error, and the transmittance in different wavelengths.

Multi-band polarization imaging provides complementary information to conventional imaging sensors, such as material components and surface characteristics. An image data in multi-band polarization imaging makes an array of 3D data cubes. Each pixel corresponds to a certain point in a high dimensional space where a statistical analysis of the data leads to object detection. There are two aspects in object detection using multi-band polarization imaging. From a theoretical background of multivariate statistical analysis, multi-band polarization image data can be considered as common multivariate data to express statistical differences between the target object and the background. Physical properties of multi-band polarization imagery can be replaced by statistical differences to make the traditional optical remote sensing processing methods useful.

Summary. Evolutionary conservation derived from a multiple sequence alignment is a powerful indicator of the functional significance of a residue, and it can help to predict active sites, ligand-binding sites, and protein interaction interfaces. The results of the existing algorithms in identifying the residue’s conservation strongly depend on the sequence alignment, making the results highly variable. Here, by introducing the amino acid similarity matrix, we propose a novel gap-treating approach by combining the evolutionary information and von Neumann entropies to compute the residue conservation scores. It is indicated through a series of tested results that the new approach is quite encouraging and promising and may become a useful tool in complementing the existing methods.

As more and more genomes have been discovered in recent years, there is an urgent need to develop a reliable method to predict the subcellular localization for the explosion of newly found proteins. However, many well-known prediction methods based on amino acid composition have problems utilizing the sequence-order information. Here, based on the concept of Chou's pseudo amino acid composition (PseAA), a new feature extraction method, the multi-scale energy (MSE) approach, is introduced to incorporate the sequence-order information. First, a protein sequence was mapped to a digital signal using the amino acid index. Then, by wavelet transform, the mapped signal was broken down into several scales in which the energy factors were calculated and further formed into an MSE feature vector. Following this, combining this MSE feature vector with amino acid composition (AA), we constructed a series of MSEPseAA feature vectors to represent the protein subcellular localization sequences. Finally, according to a new kind of normalization approach, the MSEPseAA feature vectors were normalized to form the improved MSEPseAA vectors, named as IEPseAA. Using the technique of IEPseAA, C-support vector machine (C-SVM) and three multi-class SVMs strategies, quite promising results were obtained, indicating that MSE is quite effective in reflecting the sequence-order effects and might become a useful tool for predicting the other attributes of proteins as well.

Carbon monoxide (CO) is the leading cause of death by poisoning all over the world and may result in neuropathologic changes and cognitive and neurologic sequelae, yet little is known regarding its outcomes. The present study aimed to evaluate the neuroprotective effects of N-butylphthalide (NBP) against brain damage after acute CO poisoning. The animal model of CO poisoning was established by exposed to 1000 ppm CO in air for 40 min and then to 3000 ppm for another 20 min. RT-PCR was used to assess the expressions of apoptosis-associated genes Bcl-2 mRNA and Bax mRNA. Mitochondrial membrane potential (MMP) was detected by fluorescent probe JC-1. Immunohistochemistry stain and Western blot assay were used to evaluate the expression levels of Kelch-like ECH-associated protein 1 (Keapl), nuclear factor erythroid 2-related factor 2 (Nrf-2) andquinone oxidoreductase 1(NQO-1). CO poisoning could increase the levels of Bcl-2 mRNA and Bax mRNA expressions, and obviously decrease the MMP of cells. NBP treatment could maintain the high MMP, significantly up-regulate Bcl-2 mRNA and down-regulate Bax mRNA expression, and the ratio of Bcl-2 mRNA/Bax mRNA expressions was higher than that in the CO poisoning group (P0.05). CO poisoning could start oxidative stress response. The expressions of Keap1, Nrf-2 and NQO-1 proteins significantly increased at 1, 3 and 7 day after NBP administration as compared with the CO poisoning group (P0.01). These findings suggest that N-butylphthalide may protect mitochondrial function, balance the expressions of anti-apoptosis genes and pro-apoptosis genes, be in part associated with activation of Keap1-Nrf-2/antioxidant response element (ARE) signaling pathway, and play a neuroprotective role in brain damage after acute CO poisoning.

In this study, we aim to investigate the physiochemical and biological properties of water-soluble phosphorylated polysaccharides (P-DIP) obtained from a water-insoluble polysaccharide (DIP) extracted from Dictyophora indusiata. A series of physiochemical properties were determined, including morphology, water-solubility, molecular weight, and degree of substitution (DS). To investigate the antioxidant activity of P-DIP, we determined the scavenging activity of hydroxyl radicals and DPPH, as well as the reducing power. MTT assay was performed to determine the cytotoxic effects of DIP and P-DIP on the cellular proliferation of MCF-7 and B16 cells. Compared with DIP, P-DIP showed a satisfactory water-solubility and significant increase in the antioxidant properties. Moreover, P-DIP also showed more significant inhibitory effects on the growth of MCF-7 and B16 tumor cells than the water-insoluble DIP. These results indicated that phosphorylation might contribute to the improvement of water solubility, as well as antioxidant and anti-tumor activities of natural DIP.

The dynamic kinetic resolution of alcohols (I) with 4-chlorophenyl valerate as acyl donor proceeds in the presence of Novozym 435 and sulfated TiO2 as nano solid super acidic racemization catalyst.

Porous structured regenerated cellulose films were oxidized by periodate oxidation to obtain 2,3-dialdehyde cellulose (DARC) films, which were then reacted with collagen to obtain DARC/Col composite films. The subsequent FT-IR spectra indicated that collagen was immobilized on the DARC matrix via the Schiff base reaction between NH2 in collagen and CHO in DARC backbone. Scanning electron microscopy revealed that DARC/Col exhibited a refined 3D network structure and its porosity and pore size decreased with increasing of collagen concentration. The composite films demonstrated a good equilibrium-swelling ratio, air permeability and water retention properties. The composite films also showed excellent mechanical properties, which was vital for practical application. Finally, the cytotoxicity of the composite film was evaluated using NIH3T3 mice fibroblast cells, the results revealed that DARC/Col composite films have good biocompatibility for use as scaffold material in tissue engineering.

Effect of PEG-mediated pore forming on Ca-alginate immobilization of nitrilase-producing bacteria Pseudomonas putida XY4 was studied. Through using PEG as porogen, the environmental tolerance as well as the biocatalytic reaction efficiency of immobilized cells was greatly improved, i.e., Ca-alginate-PEG immobilized cells got better temperature and substrate concentration tolerance than Ca-alginate immobilized cells and showed similar efficiency with free cells, suggesting that the intrinsic mass transfer resistance of immobilization obviously decreased. It was also observed that the pore diameter and porosity of immobilization beads were related with the molecular weight of PEG. PEG400 was found to be a relatively suitable porogen for Ca-alginate-PEG immobilized cells catalyzed hydrolysis of glycinonitrile. It was noteworthy that the Ca-alginate-PEG immobilized cells could be reused more than 18 times with little loss of enzyme activity which had shown good operation ability and great application potential.

A challenging task of airborne operations remains the landing on the carrier deck, which limits the carrier operational efficiency during rough sea. Although existing corners detection algorithm can work well without consideration of scale and orientation distortion in cooperative object, extracting corners of cooperative object at the condition of scale distortion and orientation distortion even needs more to considerate. To solve this problem, this paper presents a novel approach for UAV's automatic landing on the ship's deck. We present the design of the cooperative object and then extract the corners of cooperative object based on Scale-invariant feature transform partition bidirectional matching. Firstly, SIFT features of cooperative object are extracted from the reference image and real-time image respectively. Based on this, the strategy of partition bidirectional matching is used to eliminate the error matched points. At last, the corners of cooperative object in real-time image can be obtained through calculating the affine transformation between reference image and the real-time image. Experimental results show that the proposed algorithm is not only accurate and robust, but also feasible for real-time applications.

Automatic classification of terrain surfaces from an aerial image is essential for an autonomous unmanned aerial vehicle (UAV) landing at an unprepared site by using vision. Diverse terrain surfaces may show similar spectral properties due to the illumination and noise that easily cause poor classification performance. To address this issue, a multi-stage classification algorithm based on low-rank recovery and multi-feature fusion sparse representation is proposed. First, color moments and Gabor texture feature are extracted from training data and stacked as column vectors of a dictionary. Then we perform low-rank matrix recovery for the dictionary by using augmented Lagrange multipliers and construct a multi-stage terrain classifier. Experimental results on an aerial map database that we prepared verify the classification accuracy and robustness of the proposed method.

In order to reduce the number of flight test of the visual landing navigation verification system, reduce the cost of research and shorten the development cycle, the visualization simulation system based on virtual reality technology is demanded. An vision-based landing simulation verification system for UH based on Vega Prime / MFC was put forward. Firstly, the overall framework of the system was designed and the interaction among modules was introduced. Then a detailed description of each module function was given, and explanation of the cooperation target detection and UH position and orientation parameters were further carried out. Finally, a MFC-based Vega Prime view of the driver to achieve visualization of vision-based landing simulation verification system was developed. The simulation results show that the system can effectively unfold the cooperation target detection and UH position and orientation solving process in an intuitive, continuous and real-time pattern, and realize the simulation-testing function. It provides a good simulation platform for practical engineering applications.

Drogue detection is important for vision-based autonomous aerial refueling. It is a difficult task due to disturbances caused by both the tanker wake vortex and atmospheric turbulence. In this paper, the problem of drogue detection is considered as moving object detection. A method based on multi-scale low rank and sparse decomposition is proposed for drogue detection. Firstly, the image sequences are decomposed by stationary wavelet transform respectively. Then the method based on low rank and sparse decomposition is employed on the low frequency sub-band image sequences of coarsest scale. After obtaining the object, the object is used as object confidence map to feedback the low sub-band image sequences of next fine scale for low rank and sparse decomposition. Thereinto, an alignment method is introduced into the low rank and sparse composition due to the vibration in the drogue video image. This method can overcome the influence of non-structure information, caused by atmospheric or aircrafts. The experimental results show that the proposed algorithm is effective in real autonomous aerial refueling data via multi-scale low rank and sparse decomposition for drogue detection under complex background.

Background Protein-RNA interactions play an important role in numbers of fundamental cellular processes such as RNA splicing, transport and translation, protein synthesis and certain RNA-mediated enzymatic processes. The more knowledge of Protein-RNA recognition can not only help to understand the regulatory mechanism, the site-directed mutagenesis and regulation of RNA–protein complexes in biological systems, but also have a vitally effecting for rational drug design. Results Based on the information of spatial adjacent residues, novel feature extraction methods were proposed to predict protein-RNA interaction sites with SVM-KNN classifier. The total accuracies of spatial adjacent residue profile feature and spatial adjacent residues weighted accessibility solvent area feature are 78%, 67.07% respectively in 5-fold cross-validation test, which are 1.4%, 3.79% higher than that of sequence neighbour residue profile feature and sequence neighbour residue accessibility solvent area feature. Conclusions The results indicate that the performance of feature extraction method using the spatial adjacent information is superior to the sequence neighbour information approach. The performance of SVM-KNN classifier is little better than that of SVM. The feature extraction method of spatial adjacent information with SVM-KNN is very effective for identifying protein-RNA interaction sites and may at least play a complimentary role to the existing methods.

The microbes in the world's oceans are most abundant organisms on earth, playing an important role in the maintenance the balance of marine ecology. However, little knowledge of ecological interdependencies is known due to the limitation of current method for large-scale data and narrow surveys done for marine microbes while microbe exhibited significant inter-lineage associations naturally. Here we present a similarity network-based method to represent and analyze potential interactions among the marine microbes based on the 16S rRNA sequences. A set of parameters such as network degrees, short path, clustering coefficient and so on, are computed to characterize the similarity network topology. A few core sub networks (or network motifs) were found which show that microbe in the marine environment has a cluster propensity and evolutionary relatedness, meanwhile, the variable of network motif also indicated that the microbial diversity has a regional difference. These results show the network-based methods are effective for advance understanding the complexity and function of the marine microbial community after experiment technical.

For the instrument limitation and imperfect imaging optics, it is difficult to acquire high spatial resolution hyperspectral imagery. Low spatial resolution will result in a lot of mixed pixels and greatly degrade the detection and recognition performance, affect the related application in civil and military fields. As a powerful statistical image modeling technique, sparse representation can be utilized to analyze the hyperspectral image efficiently. Hyperspectral imagery is intrinsically sparse in spatial and spectral domains, and image super-resolution quality largely depends on whether the prior knowledge is utilized properly. In this article, we propose a novel hyperspectral imagery super-resolution method by utilizing the sparse representation and spectral mixing model. Based on the sparse representation model and hyperspectral image acquisition process model, small patches of hyperspectral observations from different wavelengths can be represented as weighted linear combinations of a small number of atoms in pre-trained dictionary. Then super-resolution is treated as a least squares problem with sparse constraints. To maintain the spectral consistency, we further introduce an adaptive regularization terms into the sparse representation framework by combining the linear spectrum mixing model. Extensive experiments validate that the proposed method achieves much better results.

In order to develop a backup navigation scheme for allowing temporally GPS faults or degradations, this paper proposes a dynamic key-frame-based natural-landmark scene matching visual navigation method for UAV. Firstly, this method could autonomously describe and check featured natural landmarks by analyzing image sequence from on-board camera. Secondly, After abstraction of key-frames including featured natural landmark, UAV will be located by the means of NLSM (Natural-Landmark Scene Matching) which based on dynamic key-frame. Thirdly, this navigation scheme adopt inter-frame scene matching algorithm in order to improving the navigation performance of accuracy, reliability and runtime. Experiments show that the vision navigation scheme proposed fits the requirements of navigation in complex and unknown environment for UAV.

In the scene matching aided-navigation system, there often exists rotation, scale difference and occlusion between actual image and reference image. According to the navigation system demands of robustness, instantaneity and precision, a multi-level algorithm based on adaptive polar transform is proposed for its robustness to rotation, scale difference and occlusion. In the matching process, the rectangle window is replaced by a circle window to eliminate the cut error caused by rotation, which improves the matching precision. To accelerate the matching procedure, a pixel-jump searching method combined with a multi-resolution imaging technique is used to reduce the computational load. Experiments illustrate the effectiveness of the proposed methods.

Protein-RNA interactions are vitally important to a number of fundamental cellular processes, including regulation of gene expression such as RNA splicing, transport and translation, protein synthesis and assembly of ribosome. More detailed information on the Protein-RNA interaction is helpful for comprehending the function notation and molecular regulatory mechanism, meanwhile, knowing the knowledge of Protein-RNA recognition can also help the biological scientist and researcher understand the site-directed mutagenesis and drug design. In the present work, we proposed a computational approach, based on SVM-KNN algorithm, with evolutionary information of spatial neighbour residues for prediction of protein-RNA interaction sites. The overall success rate obtained by 5-fold cross-validation is 78.00%, which is comparable or better than other existing methods, indicating our method is very promising for identifying and predicting protein-RNA interaction sites.

DS evidence theory is a hot topic in multi-sensor data fusion because of its advantages in multi-source information presentation and processing. Considering the difficulties of sensor reliability priority information acquisition, how to compute reliability of sensor via the evidence distance in multi-sensor data fusion system and how to select consistent evidence for combination via the evidence distance when many homogeneous sensors exist are the open issues. In this paper, a new evidence distance is presented in order to avoid the disadvantage of Diaz's distance measure in the logical selection of tuning parameter and Jousselme's distance measure in not considering the proximity of reference frame to the frame of discernment. This method firstly defines a fuzzy linguistic term set to express the proximity of reference frame to the frame of discernment, in which the semantics of the terms is given by a triangular membership function and a rewarding or penalizing function. Then similarity measurement matrix is achieved by weighting the rewarding or penalizing function using membership degree. Finally Jousselme's distance framework is used to compute the evidence distance. Two numerical examples illustrate the efficiency of the method proposed.

In this paper, a rotation invariant multi-model scene matching method is proposed for scene matching aided navigation system. Phase congruency transformation is introduced first to minimize the image difference between multi-model images. Then ring project transformation is processed to make the method invariant to rotation. However, multiple maximum phenomenon is likely to occur after ring project transformation. To solve this problem, a multi-frame spatial-temporal correlation matching method is proposed. Using the spatial-temporal correlation gained from the inertia system or matching of the adjacent inter-frames, an optimal matching position is gained by maximizing a multi-correlation surface. Afterward, surface fitting method is used to get sub-pixel accuracy matching position. This method, which is invariant to rotation, greatly increases match accuracy. Necessary simulation proves the efficiency of our method.

A combination of Novozym435 with an acidic polystyrene resin CD8604 affords a new and efficient dynamic kinetic resolution of acetophenones to give the corresponding benzylic alcohols with good to excellent enantioselectivities.

Acoustic source localization is a very important Wireless Sensor Network (WSN) surveillance task. In real-world implementations, the localization algorithm, which is essentially an optimization of a certain cost function involving all received sensor observations, must be feasible under stringent communication, computation and energy constrains. In this paper, we propose a novel light-weight dynamic population Particle Swarm Optimization (PSO) based method to search the Maximum Likelihood Estimate (MLE) solution for the source location, reducing computation complexity as well as mitigating the affects of local optimum solutions. Moreover, instead of raw data, only signal energy observations are transmitted to the fusion center, such that bandwidth and energy consumptions can be largely decreased. The results of extensive simulations have shown the superior performance of our method in various scenarios.

The classical Dempster-Shafer theory involves counter-intuitive behaviors when the evidence high conflict. In order to solve the problem, a new approach of weighted evidence combination is proposed. Both evidence distance and inconsistent information are considered to evaluate the conflict. Each piece of evidence is given weight coefficient according to its amount of conflict with others. Then two methods are used to modify the belief functions, and the modified functions are combined by DS rule. Numerical example shows the proposed approach can efficiently handle conflicting evidence with good convergence efficiency.

Video captured from Unmanned Aerial Vehicles (UAV) usually suffer from undesired motion of the sensors, which severely defects detecting and tracking targets of interest. In this paper, a fast video stabilization algorithm for UAV is proposed to remove undesired motions and produce stabilized images. Firstly, a set of fast matching strategies are considered to match successive frames. Then, with motion parameters between two successive frames, a polynomial fitting and predicting method is proposed to estimate the global motion parameters and find out the undesired frames. After that, the undesired images are compensated. Finally, all frames are transformed to obtain stabilized images. Experimental results demonstrate the stabilization and the efficiency of the proposed algorithm.

The problem of infrared targets detection in heavy clutter background is unsolved well. In this paper, we proposed a matched filter method based on tensor model, which generalizes the traditional matched filter, is used to distinguish between targets and clutter using multispectral infrared image. The tensor matched filter approach is adopted to take into account both spatial and spectral information simultaneously. The 3-mode product between tensor and filter matrix is given for matched filter. This method doesn't destroy the original spatial structure of multispectral images. By considering the spatial and spectral information for the filter, the probability of detection is improved. And an improved multi-way filter for multispectral tensor is performed before tensor matched filter. The experimental results of ROC curves indicate that our algorithm is effective to improve detection performance of targets in clutter background. And the multispectral images acquired at 8-9 μm , 9-10 μm , 10-11 μm , 11-12 μm , and 8-14 μm were tested in the experiment. We also show that this novel approach is extremely efficient in applications of interest by using tensor model for multispectral IR image processing.