Rapid and efficient hand gestures recognizer based on classes discriminator wavelet networks

Multimedia Tools and Applications - The vision based on hand gesture recognition is one of the key challenges in behavior understanding and computer vision. It offers to machines the possibility of...     

A multiresolution wavelet networks architecture and its application to pattern recognition

This paper aims at addressing a challenging research in both fields of the wavelet neural network theory and the pattern recognition. A novel architecture of the wavelet network based on the multiresolution analysis (MRWN) and a novel learning algorithm founded on the Fast Wavelet Transform (FWTLA) are proposed. FWTLA has numerous positive sides compared to the already existing algorithms. By exploiting this algorithm to learn the MRWN, we suggest a pattern recognition system (FWNPR). We show firstly its classification efficiency on many known benchmarks and then in many applications in the field of the pattern recognition. Extensive empirical experiments are performed to compare the proposed methods with other approaches.     

Nonlinear state feedback design for continuous polynomial systems

The objective of this paper is to simplify the complexity of practical implementation of the input-state feedback linearization technique for the control of input-affine systems. A polynomial approach which makes use of the Taylor series expansion and the Kronecker product is developed. Our work aims to address the problem of synthesizing a polynomial control via a nonlinear analytical coordinates transformation. To check the effectiveness of the investigated approach, we consider the control problem of a series DC motor. A comparative study with the input-state feedback linearization control is developed.     

Improving the performance of retail stores subject to execution errors: coordination versus RFID technology

This paper analyzes a Newsvendor type inventory model in which a manufacturer sells a single product to a retailer store whose inventory is subject to errors stemming from execution problems. Hence, within the store, all of the products are not available on shelf for sales either because the replenishment of the shelf from the backroom is subject to execution errors that mainly result in products lost in the backroom or products misplaced on the other shelves of the store. We compare two situations: in the first situation, the two supply chain actors are aware of errors and optimize their ordering decisions by taking into account this issue. The second situation deals with the case where an advanced automatic identification system such as the Radio Frequency Identification technology is deployed in order to eliminate errors. Each situation is developed for three scenarios: in the centralized scenario, we consider a single decision-maker who is concerned with maximizing the entire supply chain’s profit; in the decentralized uncoordinated scenario, the retailer and the manufacturer act as different parties and do not cooperate. The third scenario is the decentralized coordinated scenario where we give conditions for coordinating the channel by designing a buyback contract. The authors would like to thank the editor, Prof. Dr. Hans-Otto Guenther, and two anonymous referees for their constructive comments to improve this article.     

The news-vendor problem with drop-shipping and resalable returns

As e-commerce expands, more and more products are offered online to attract internet consumers’ interest. These products are then shipped to consumers’ home by a drop-shipper. Drop-shipping seems to be a good option to sell products in addition to physical stores. Furthermore, both types of products, either sold in store or on Internet can be returned by consumers, with often a higher return ratio for those purchased on Internet. To model these two sales channel and the interactions between them, we consider a News-Vendor (NV) managing both a physical store and an online sale channel that can be fulfilled by a drop-shipping option. We also consider the possibility of reselling products that are returned by consumers during the selling season. The concavity of the expected profit is proven and the optimality condition is obtained. Promising results are obtained from a numerical analysis. In particular, we show that the expected profit can be 14.4% less than the optimal expected profit if the return effect is ignored. Using drop-shipping option can reduce the optimal store inventory by 31.2% and if the NV has no drop-shipping option, the expected profit can be 9% less.     

Effects of Transition-Metal V-Doping on the Structural, Magnetic and Transport Properties in La0.67Sr0.33MnO3 Manganite Oxide

We have investigated the structural, magnetic, and electrical transport properties of a series of ABO₃-type perovskite compounds, La_0.67Sr_0.33Mn_1?x V _x O₃ (0≤x≤0.15). The samples were characterized by X-ray diffraction and data were analyzed using Rietveld refinement technique, it has been concluded that these materials have the rhombohedral structure with $\mathrm{R}\overline{3}\mathrm{C}$ space group. The magnetization and resistivity measurements versus temperature proved that all our samples exhibit a ferromagnetic to paramagnetic transition and a metallic to semiconductor one when the temperature increases. Both the Curie temperature T _C and the resistivity transition temperature T _P of the composites decrease, while the resistance increases as the V content increases. It has been concluded that the electrical conduction mechanism in the metallic regime at low temperatures (T<T _P) can be explained on the basis of grain boundary effects and the single electron-magnon scattering process. Resistivity data were well fitted with the relation ρ=ρ ₀+ρ ₂ T ²+ρ _4.5 T ^4.5, whereas the adiabatic Small Polaron Hopping (SPH) and Variable Range Hopping (VRH) models are found to fit well in the paramagnetic semiconducting regime at the high temperature (T>T _P).     

Towards Securing Machine Learning Models Against Membership Inference Attacks

From fraud detection to speech recognition, including price prediction, Machine Learning (ML) applications are manifold and can significantly improve different areas. Nevertheless, machine learning models are vulnerable and are exposed to different security and privacy attacks. Hence, these issues should be addressed while using ML models to preserve the security and privacy of the data used. There is a need to secure ML models, especially in the training phase to preserve the privacy of the training datasets and to minimise the information leakage. In this paper, we present an overview of ML threats and vulnerabilities, and we highlight current progress in the research works proposing defence techniques against ML security and privacy attacks. The relevant background for the different attacks occurring in both the training and testing/inferring phases is introduced before presenting a detailed overview of Membership Inference Attacks (MIA) and the related countermeasures. In this paper, we introduce a countermeasure against membership inference attacks (MIA) on Conventional Neural Networks (CNN) based on dropout and L2 regularization. Through experimental analysis, we demonstrate that this defence technique can mitigate the risks of MIA attacks while ensuring an acceptable accuracy of the model. Indeed, using CNN model training on two datasets CIFAR-10 and CIFAR-100, we empirically verify the ability of our defence strategy to decrease the impact of MIA on our model and we compare results of five different classifiers. Moreover, we present a solution to achieve a trade-off between the performance of the model and the mitigation of MIA attack.     

Nature of disorder in synthetic hectorite

The disorder of synthetic Na-hectorite was investigated using structural simulation of X-ray powder patterns.The method used to characterize this clay mineral was based on the comparison between experimental and calculated X-ray diffraction profiles. The study was achieved in two steps: (i) the quantitative interpretation of 00l reflections enabled the determination of the number of intercalated water molecules, their positions and the stacking mode of the clay mineral layers along the normal to the (a, b) plane; and (ii) the study of the hkl reflections with h and/or k ≠ 0 enabled the characterization of the structural evolution in the (a, b) plane of synthetic hectorite. Many models of disorder with random shifts of ± a/3, a/3 + b/3, 80% of a/3 + b/3 and 20% of a/3 with a total density of defects of 80%, gave the best fit to the observed data at the level of (02, 11), (20, 13), (04, 22), (15, 24, 31) and (06, 33) bands with 5° of tetrahedral cavity distortion (the rotation angle of the tetrahedra).     

Effect of storage on refined and husk olive oils composition: Stabilization by addition of natural antioxidants from Chemlali olive leaves

Two samples of refined olive and husk oils have been analysed in order to evaluate the influence of storage time on their quality. The following parameters were determined: peroxide values, absorption coefficients K270 and K232, Rancimat induction time, sterols and fatty acid contents. Six months storage at 50 °C in the dark revealed a loss in oil stability. This finding was reflected by the greater increase in peroxide value and a decrease of Rancimat induction time and sterol content. The enrichment of refined olive and husk oils with olive leaves and its hydrolysate extract resulted in an appreciable resistance to oxidative deterioration due to its phenolic antioxidants content. Oleuropein and hydroxytyrosol were the major compounds in Chemlali olive leaves extract and hydrolysate solution, respectively. The antiradical activity of leaves extract as well as its hydrolysate solution was evaluated and compared to that of the BHT. The antioxidant activity of the enriched refined olive and husk oils with leaves and hydrolysate extracts at 400 ppm showed that the latter had the highest protective effect against oil oxidation. Oils with added hydrolysate extract had the lower peroxide value and the higher stability measured with a Rancimat method. After six months of storage the induction time increased from 23.3 to 83.5 h for refined olive oil and from 16.6 to 49 h for husk oil. Furthermore, during oil storage, there was no significant variation in fatty acid composition. However, the total sterol concentration of the oils treated with hydrolysate extract increased. The results suggested that hydrolysate and leaves extracts are excellent antioxidants and can serve as substitutes for synthetic antioxidants.