Search

The quaternary orthogonal moments have been widely used as color image descriptors owe to their remarkable color and shape information encapsulation capability. Their computation, however, depends on finding the optimal value of a unit pure quaternion parameter, which is done empirically and with no warranty of optimality. We propose a 2D color object recognition method that relies on the quaternion-valued parameter-free disc-harmonic moment invariants (QHMs) fed into the quaternion extreme learning machine (QELM). The role of this latter is to maintain the correlation between the four parts, real and imaginary, of the quaternary descriptor coefficients. Several datasets are used for recognition experiments. We draw the conclusion that: (1) our quaternion-valued QHMs invariants outperform other quaternary moments, (2) the quaternion-valued moment invariants give results better than the modulus-based moment invariants and (3) the QELM yields results better than the state-of-the-art classifiers.

Extractive multi-document summarization (MDS) is the process of automatically summarizing a collection of documents by ranking sentences according to their importance and informativeness. Text representation is a fundamental process that affects the effectiveness of many text summarization methods. Word embedding representations have shown to be effective for several Natural Language Processing (NLP) tasks including Automatic Text Summarization (ATS). However, most of these representations do not consider the order and the semantic relationships between words in a sentence. This does not fully allow grasping the sentence semantics and the syntactic relationships between sentences constituents. In this paper, to overcome this problem, we propose an unsupervised method for generic extractive multi-document summarization based on the sentence embedding representations and the centroid approach. The proposed method selects relevant sentences according to the final score obtained by combining three scores: sentence content relevance, sentence novelty, and sentence position scores. The sentence content relevance score is computed as the cosine similarity between the centroid embedding vector of the cluster of documents and the sentence embedding vectors. The sentence novelty metric is explicitly adopted to deal with redundancy. The sentence position metric assumes that the first sentences of a document are more relevant to the summary, and it assigns high scores to these sentences. Moreover, this paper provides a comparative analysis of nine sentence embedding models used to represent sentences as dense vectors in a low dimensional vector space in the context of extractive multi-document summarization. Experiments are performed on the standard DUC’2002–2004 benchmark datasets and the Multi-News dataset. The overall obtained results have shown that our method outperforms several state-of-the-art methods and achieves promising results compared to the best performing methods including supervised deep learning based methods.

Vehicular Ad-hoc Networks (VANETs) have been suggested as an active and powerful field of research to mitigate environmental problems and challenges. The main challenge in a VANET is to ensure routing with a good Quality of Service (QoS). The Greedy Perimeter Stateless Routing (GPSR) protocol is one of the most promising position-based routing mechanisms used to overcome this challenge. Its effectiveness depends entirely on the information on a node's mobility and the precision of this information. By broadcasting periodic beaconing within transmission boundary ranges, GPSR can manage neighbors' mobility information and maintain up-to-date lists of neighbours. Nevertheless, information on the position of a neighboring vehicle quickly becomes outdated, which negatively influences the efficiency of the routing. In order to monitor information mobility and to increase the QoS in this challenging area, position estimation needs to be considered. Thus, in this study, we examine the position estimation problem, and propose an improvement to the GPSR protocol, named KF-GPSR, where each vehicle estimates in real time the position of its neighbors using the Kalman filter algorithm. Indeed, by employing this strong estimation technique, it is possible to reduce considerably the frequency of exchanged beacon packets, while maintaining high position accuracy. For greater reliability, we also propose an extension to KF-GPSR, called BOD-KF-GPSR, that uses the “beacon-on-demand” process only if a node needs to rediscover its neighborhood. Simulation experiments using the network simulator NS-2 are presented to demonstrate the ability and usefulness of our two proposals. Here, we compare the proposed protocols against diverse common protocols: GPSR, AODV, DSR, and ZRP. The results show that BOD-KF-GPSR achieves a significant enhancement in terms of its packet delivery ratio, routing cost, normalized routing load, end-to-end delay, and throughput.

International audience; Term mismatch is a common limitation of traditional information retrieval (IR) models where relevance scores are estimated based on exact matching of documents and queries. Typically, good IR model should consider distinct but semantically similar words in the matching process. In this paper, we propose a method to incorporate word embedding (WE) semantic similarities into existing probabilistic IR models for Arabic in order to deal with term mismatch. Experiments are performed on the standard Arabic TREC collection using three neural word embedding models. The results show that extending the existing IR models improves significantly baseline bag-of-words models. Although the proposed extensions significantly outperform their baseline bag-of-words, the difference between the evaluated neural word embedding models is not statistically significant. Moreover, the overall comparison results show that our extensions significantly improve the Arabic WordNet based semantic indexing approach and three recent WE-based IR language models.

Vehicular ad-hoc networks (VANETs) are still more challenging to overcome even if they have been widely studied during the last decades. The routing mechanism is the essentially relevant issue in this field. Indeed, it must strictly to be adapted to specific and unique characteristics such as the high mobility of the vehicles, the dynamic nature of network topology as well as the high link breakage probability. In this paper, our objective is to improve the greedy perimeter stateless routing protocol (GPSR) as being the most promising position-based mechanism. However, according to the impact of position information on routing decision, our proposed approach defined by GPSR+PRedict protocol ensures that each vehicle estimates its own position for the near future. Afterwards, through extensive experiments, we show the ability of the GPSR+PRedict to overcome the observed problems and to enhance the overall performance of the traditional GPSR approach. The simulations are carried out on both highway and urban scenarios by using NS-2 and VanetMobiSim simulators.

Unlike French and English, the richness and ambiguity of written Arabic texts cause a great deal of errors. The purpose of this article is to resolve issues of tolerance of some errors in Arabic texts and to develop an automatic detection system as well as a correction system of those errors. This work represents a combination of the Levenshtein Distance (LD) and bi-context language models based on a large training corpus. The method involves the automatic detection and correction of no-word1 errors and the real words2 errors. This method consists of two parts. The first part is to extract the context information from the training corpus, which means to create the vocabulary and the database of the bi-context language models. The second part implements the automatic detection and correction of the incorrect words. The experimental results show a successful extraction of the context information from the corpus. Furthermore, they also show that the use of this system can reduce the rate of errors in Arabic texts.

In vehicular ad-hoc network (VANET), routing was and will remain an active field for research as long as the researchers can improve it. In fact, VANET encounters many confrontations due to the high speed and mobility pattern. Thus, the most dialectical and debatable issue is how to ensure the routing in a brief delay and with a minimum lost packets. To this end, various geographical routing protocols have been proposed to confront many challenges. One of these protocols is the greedy perimeter stateless routing (GPSR) that represents the most promising position-based routing protocol in wireless networks. The nodes use the current location information of the neighbours within their transmission boundary range and the location of the destination. These information are important to choose the adequate next hop for forwarding data. However, there are some cases when this strategy of GPSR is not suitable. The paper presents an enhancement of this famous protocol identified as GPRS+PRedict. This improvement is defined by estimating the future position of all participating nodes. Our experimental results show the ability of the GPSR+PRedict protocol to overcome the observed problem and to improve the overall performance of the traditional GPSR. The simulation is carried out on highway scenario by using NS-2 and VanetMobiSim simulators.

Stemming is a process of reducing inflected words to their stem or root from a generally written word form. This process is used in many text mining application as a feature selection technique. Moreover, Arabic text summarization has increasingly become an important task in natural language processing area (NLP). Therefore, the aim of this paper is to evaluate the impact of three different Arabic stemmers (i.e. Khoja, Larekey and Alkhalil's stemmer) on the text summarization performance for Arabic language. The evaluation of the proposed system, with the three different stemmers and without stemming, on the dataset used shows that the best performance was achieved by Khoja stemmer in term of recall, precision and F1-measure. The evaluation also shows that the performances of the proposed system are significantly improved by applying the stemming process in the pre-processing stage.

In order to describe color images, the use of the algebra of quaternions in combination with existing image orthogonal moments, meant for binary and grayscale images, has been widely investigated. This is because of their advantages in (1) gathering the three-channel color information in a single feature vector while preserving the correlation between them and in (2) eliminating shape information redundancy. However, the computation of these quaternary orthogonal moments depends on a unit pure quaternion parameter. The optimal value of this latter can be fixed only with the help of experiments and it is application-dependent. We propose a parameter-free formulation of the quaternary orthogonal moments. The general formula for the computation of the proposed moments, whose rotation invariance is achieved by retaining the modulus, is provided. Furthermore, experiments are conducted to evaluate the performance of the proposed modulus-based moment invariants for color image retrieval and recognition.

We introduce a hybrid fingerprint recognition method built from minutiae and quaternion orthogonal moments. The proposed algorithm includes four steps: extraction of the minutiae triplets (m-triplets), first pass of triplets minutiae matching, validation step of these triplets by characterizing their neighboring gray-level image information through feature vectors of quaternion radial moments, and an adequate similarity measure. By boosting the local minutiae matching step, we avoid consolidation and global matching. To show the added-value of our method, several algorithms for extracting and matching m-triplets are considered and an experimental comparison is established. Experiments are carried out using all four parts of the FVC2004 dataset. Results indicate that the combination of the geometrical features and the quaternion radial moments of the m-triplets leads to an improvement in the overall fingerprint matching performance and demonstrate the expected gain of integrating a validation step in an m-triplets based fingerprint matching algorithm.

A set of invariant quaternion moments based on an adaptation of the three-dimensional (3-D) spherical harmonic transform (SHT) for describing two-dimensional color shapes is proposed. The use of quaternions to deal with the color part is beneficial in the way the three color components are integrated in a single feature. An adequate mapping from the 3-D SHT to the unit disc allows a fast and accurate computation of the proposed moments. Experiments are conducted to evaluate the performance of the obtained moments in terms of color image reconstruction, robustness to geometric and photometric transformations, content-based color shape retrieval, and computation time. For this purpose, two image databases (COIL-100 and ALOI) are used. Results illustrate the effectiveness of the proposed moments in dealing with the color information.