Advantage of the increasing resolution of radar systems is the opportunity to have more details characteristic of a specific **target**. Disadvantage is that these detailed characteristics require more and more computer memory to be stored, computer resources and increase the search computational time to NCTR (**Non**-**Cooperative** **Target** **Recognition**). It is therefore important to develop efficient methods to decrease the size of high resolution data of radar targets. One way to compress these data is to use tree structured representation using clustering algorithm coupled with a multiresolution wavelet representation to decrease the data size and the number of RCS signature [1].

En savoir plus
. (10)
6.2 Application of Multiresolution Hierarchical Tree to **Target** **Recognition**
Different multiresolution hierarchical trees have been designed from different beginning decomposition levels (1 to 4). An example of tree built from the decomposition level 4, and using the Haar wavelet and the K-means hard partitioning algorithm, is shown in Figure 7. This tree has 21 final clusters, an average distortion of 0.56 and a partition entropy of 2.9. In this figure, the clusters are designated by the notation C j k , , where k is the cluster number at resolution j. The number in each circle defines the percentage of data in the cluster.

En savoir plus
Abstract—In this paper, the problem of efficient
representation of large database of **target** radar cross section is investigated in order to minimize memory requirements and **recognition** search time, using wavelet representation. Synthetic RCS of large aircrafts, in the HF-VHF frequency bands, are used as experimental data. Many parameters are evaluated like mother wavelet, decomposition level, and classification parameters. Criteria used to determine the efficiency of multiresolution representations are compression scores, false identification rate and search time.

In order to initialize the camera pose for Object Tracking, 3D-2D correspondences will be computed later. These can be used to compute the camera pose w.r.t. the object. Since feature matching is performed in the images, we need to assign 3D coordinates to every feature. We do so using a specifically dedicated graphical tool to Assign 3D-Coordinates. As depicted in Figure 3a, the data-flow between the 3 components of the off-line training step is straightforward. An image of the object is the only input of the system. The location of the features is an extra input for the description phase. In order to compute the 3D coordinates of the features, a (simplified) 3D model of the object is needed as well. During real operation, the system needs to identify objects in the image or certify their presence. This is the goal of the object **recognition** phase, implemented in the Object **Recognition** activity. The data flow diagram of this activity is shown in Figure 3b. We recognize the first two components of this phase. The Feature Extraction and Feature Description components are identical to those in the off-line training phase. Indeed, the first step in recognizing an object in an image consists of locating features in this image and describing these features using the same algorithm as before. The newly found feature descriptors can then be matched to the feature descriptors of the objects in the database. This is done in the Feature Matching component. The result of this component consists of matches between features, i.e. 2D-2D correspondences. These results can contain mismatches, while other (correct) matches might have been missed. This can be ameliorated by the Verification component, which will output its result in the form of matches between 3D coordinates (found by the Assign 3D Coordinates component in the pre-processing phase) and 2D coordinates (of the features extracted in the **target** image. These 3D-2D correspondences can be used to compute an initialization of the camera pose w.r.t. the object.

En savoir plus
Keywords: Interpersonal stance, **non**-verbal behavior in- terpretation, Social Signal Processing
1 Introduction
The last two decades have seen a surge of in- terest in the field of Human-Computer Interac- tion for the introduction of Embodied Conversa- tional Agents (ECA) in various application do- mains, such as interactive storytelling [8], vir- tual learning environments [17], healthy beha- viour promotion [7], or museum guides[13]. One of the major reasons behind this strong movement is that some studies found that using ECAs improved the experience of human- computer interaction, by making learning activi- ties easier to follow [17] or by enhancing the de- gree of trust users had in relationship with their computer [21].

En savoir plus
externalities. In this family, games differ by the size of maximum coali- tion, partitions and by coalition structure formation rules.
A result of every game consists of partition of players into coali- tions and a payoff profile for every player. Every game in the family has an equilibrium in mixed strategies with possibly more than one coalition. The results of the game differ from those conventionally discussed in **cooperative** game theory, e.g. the Shapley value, strong Nash, coalition-proof equilibrium, core, kernel, nucleolus.

Unité de recherche INRIA Rennes : IRISA, Campus universitaire de Beaulieu - 35042 Rennes Cedex (France) Unité de recherche INRIA Rocquencourt : Domaine de Voluceau - Rocquencourt - BP 10[r]

The tuning hyperparameter is chosen to be the scale factor σ in (2) of local kernel k. Although matching based kernel leads to a good performance, the Mercer condition in not usually insured. Fig. 2 presents ranked eigenvalues of the Gram matrix for different values of σ. For σ = 100, there exists negative eigenvalues of the Gram matrix, this means that matching kernel is not positive definite in general. By decreasing the scale σ, eigenvalues become always positive. Fig. 3-a shows the variation of the **recognition** error with respect to log 10 (σ). The optimal value obtained by cross-validation is σ ∗ = 10 −3

En savoir plus
For catalysts prepared by the Solution Combustion Synthesis (SCS), higher CO conversion values have been reached with ternary and binary oxides rather than with pure ceria, [r]

Keywords—Automatic **target** **recognition**, synthetic aperture radar, SIFT, saliency attention model, matching
I. I NTRODUCTION
Automatic **target** **recognition** (ATR) using Synthetic aper- ture radar (SAR) images has become an essential research topic for several application fields such as military defense. The ATR-SAR task aims to recognize in automatic way the unknown targets based on its SAR images. For achieving the **recognition** task, several steps are usually required including data acquisition, feature extraction and classification to build decision making [1, 2]. In the first stage, the SAR images are constructed. It is followed by feature extraction that consists of calculating a signature from each **target** image. Finally in last stage, these feature vectors are used in classification step to recognize unknown targets. We focus in this paper on the feature extraction and classification steps.

En savoir plus
where F is a column vector with components f i , 1 ≤ i ≤ n. Matrix A is not necessarily **cooperative**, that means that its terms outside the diagonal are not necessarily positive. First we introduce some notations concerning matrices. Then, with these notations we can state our results and prove them.

each player’s behavior, which is typically based on parameters like level of foresight, risk aversion, or knowledge of the other players’ preferences ( Madani and Hipel , 2011 ), which may all be difficult to assess. Moreover, while NCGT produces insightful information into strategic behaviors for studying negotiations, for instance, the results remain most of the time qualitative ( Madani et al. , 2014 ). **Cooperative** game theory (CGT) assumes that parties are already bound, and communicate and ex- change information before the game. Decisions are not taken unilaterally but jointly to lead to a Pareto equilibrium. The objective is to address the allocation problem, particularly for water-scarce basins, by developing functional water allocation arrangements ( Parrachino et al. , 2006 ; Madani et al. , 2014 ). The approach is quantitative and it is aimed at assessing the value of cooperation under different coali- tions in transboundary river basins including the Syr Darya basin between Kyrgyzstan, Uzbekistan, and Kazakhstan ( McKinney and Teasley , 2007 ; Teasley and McKinney , 2011 ); the Nile basin between Egypt, Sudan, Ethiopia, Uganda, Kenya, Tanzania, Burundi, Rwanda, Democratic Republic of Congo, and Eritrea ( Wu and Whittington , 2006 ); the Tigris and Euphrates river basins between Turkey, Syria, and Irak ( Kucukmehmetoglu and Guldmann , 2004 ); but also in western Middle East for a regional water trade between Egypt, Israel, the West Bank, and the Gaza Strip ( Dinar and Wolf , 1994 ). Stud- ies that used CGT have then worked on resource allocation methods so that the sharing can be more acceptable and produce more benefits to the riparian countries. Typical allocation systems include: (i) social welfare maximization – maximization of the basin-wide benefits as described in Eq. 1.1 and the previous section, (ii) bankruptcy methods – for fairly sharing a scarce resource between users, e.g. with the Caspian Sea ( Sheikhmohammady and Madani , 2008 ) or the Tigris River between Turkey, Syria and Iraq ( Mianabadi et al. , 2015 ), and (iii) benefit sharing – for equitably sharing the benefits through water allocations, e.g. on the Eastern Nile river between Ethiopia, South Sudan, Sudan, and Egypt ( Arjoon et al. , 2016 ).

En savoir plus
115 En savoir plus

In this article, we take our previous analysis [6] a step far- ther and investigate deeper the impact of rhythm variation sep- arately on **target** and **non**-**target** comparisons. First, we pro- pose to analyze whether some rhythmic parameters are de- pendent to the speaker. Second, we investigate if variation in rhythm may explain the high intra-speaker variability observed for some speakers and therefore explain the difference in per- formance observed between speakers. Our study is performed based on Fabiole [17], a database where within-speaker vari- ability is strong.

An alternative approach is based on autonomous, self- interested agents [5]. Such routing schemes are also known as ”selfish routing” since each dispatcher inde- pendently seeks to optimize the performance perceived by the jobs it routes. This setting can be analysed within the framework of a **non**-**cooperative** routing game. The strategy that rational agents will choose under these circumstances is called a Nash Equilibrium and it is such that a unilateral deviation will not help any routing agent in improving the performance perceived by the traffic it routes.

In this work, we introduce the logics E(G) for reasoning about probabilistic expectation over classes G of games with discrete polynomial payoﬀ functions represented by ﬁnite-valued Luk[r]

Unité de recherche INRIA Sophia Antipolis 2004, route des Lucioles - BP 93 - 06902 Sophia Antipolis Cedex France Unité de recherche INRIA Futurs : Parc Club Orsay Université - ZAC des Vi[r]

Regime of the Nuclear-Test-Ban Treaty . . . . . . . . . . . . . . . . . . . . . . . . 131
Sabine Bauer and Cormac O’Reilly
7 The Nuclear Safeguards Regime of EURATOM: A Regional Cornerstone of the Verification of **Non**-Proliferation

data with each other and then forward the signals to destination using **non**-regenerative DSTC is proposed to improve the performance of the system in terms of BER. To the best of our knowledge, this is the first time the data exchange between relays in **cooperative** relaying networks is presented and analyzed. The lower bound expression on average symbol error probability (ASEP) of full DSTC **cooperative** relaying system is derived. Beside, the upper bound expression on ASEP is given in the case that the distance between relays,

signals to far users by using significant transmission power. Therefore, energy harvesting has attracted much attention from the research community. Two practical designs, namely, time switching and power splitting, have been put forward in [2] for simultaneous power and information transmission. These two schemes have been widely accepted and used. For example, the power splitting scheme was applied to **cooperative** NOMA in [3]. It was found that simultaneous wireless information and power transfer (SWIPT) will not jeopardize NOMA’s diversity gain, and the benefit of user selection based on node locations was demonstrated. In addition, a generalized scheme combining time switching and power splitting was proposed in [4].

En savoir plus
provide good relative position accuracy at distances up to several kilometres as well as relative orientation cues at shorter distance, but are typically very expensive and power hungry, and often characterized by a small field of view (FOV). The algorithm in Ref. [5] relies on the iterative recursive least squares method (IRLS) to reject outliers proposed in Ref. [9] and includes a Kalman enhancement (as suggested by Ref. [10]) in order to improve the robustness of the estimation. The application of a **non**-linear version of RAPiD in space rendezvous has been assessed also in Refs. [11]–[13], where a graphic process units (GPU) is used in order to render not only geometrical edges but also texture discontinuities of the a priori model. Other RAPID-like methods have been proposed in Refs. [14], [15]. All the cited works integrate, to the pose estimation algorithm, a linear Kalman filter (KF) which propagates a simple kinematic model. However, in the case of high rotation rates typical of a tumbling object, a simple kinematic filter doesn’t allow to estimate the rotation rate of the **target**, which needs to be known to perform some RDV operations requiring the synchronization of chaser motion with **target** motion.

En savoir plus