Challenge For Various Applications Research Articles

Heterogeneous self-supervised interest point matching for multi-modal remote sensing image registration

ABSTRACT The implementation of multi-modal remote sensing image registration is still a challenge in various applications. In this paper, a novel framework called Heterogeneous SuperPoint Network is proposed for multi-modal remote sensing image registration. Heterogeneous SuperPoint Network operates on full-sized images and produces multi-modal interest points accompanied with fixed length descriptors in a single forward pass. Moreover, we develop a strategy for training a base detector that used in conjunction with heterogeneous homographic adaptation to generate pseudo-ground truth interest point labels for unlabeled images in a self-supervised fashion. Then, the interest points and point correspondences obtained by Heterogeneous SuperPoint Network are used to do the final matching with the nearest neighbor (NN) matching and Random sample consensus (RANSAC). Various image sets have been considered in this paper for the evaluation of the proposed approach. Experimental results demonstrate the superior performance of the proposed method over the existing design for the whole data set.

Efficient Pairing in Unknown Environments: Minimal Observations and TSP-Based Optimization

Generating paired sequences with maximal compatibility from a given set is one of the most important challenges in various applications, including information and communication technologies. However, the number of possible pairings explodes in a double factorial order as a function of the number of entities, manifesting the difficulties of finding the optimal pairing that maximizes the overall reward. In the meantime, in real-world systems, such as user pairing in non-orthogonal multiple access (NOMA), pairing often needs to be conducted at high speed in dynamically changing environments; hence, efficient recognition of the environment and finding high reward pairings are highly demanded. In this paper, we demonstrate an efficient pairing algorithm to recognize compatibilities among elements as well as to find a pairing that yields a high total compatibility. The proposed pairing strategy consists of two phases. The first is the observation phase, where compatibility information among elements is obtained by only observing the sum of rewards. We show an efficient strategy that allows obtaining all compatibility information with minimal observations. The minimum number of observations under these conditions is also discussed, along with its mathematical proof. The second is the combination phase, by which a pairing with a large total reward is determined heuristically. We transform the pairing problem into a traveling salesman problem (TSP) in a three-layer graph structure, which we call Pairing-TSP. We demonstrate heuristic algorithms in solving the Pairing-TSP efficiently. This research is expected to be utilized in real-world applications such as NOMA, social networks, among others.

A Precisely One-Step Registration Methodology for Optical Imagery and LiDAR Data Using Virtual Point Primitives

The registration of optical imagery and 3D Light Detection and Ranging (LiDAR) point data continues to be a challenge for various applications in photogrammetry and remote sensing. In this paper, the framework employs a new registration primitive called virtual point (VP) that can be generated from the linear features within a LiDAR dataset including straight lines (SL) and curved lines (CL). By using an auxiliary parameter (λ), it is easy to take advantage of the accurate and fast calculation of the one-step registration transformation model. The transformation model parameters and λs can be calculated simultaneously by applying the least square method recursively. In urban areas, there are many buildings with different shapes. Therefore, the boundaries of buildings provide a large number of SL and CL features and selecting properly linear features and transforming into VPs can reduce the errors caused by the semi-discrete random characteristics of the LiDAR points. According to the result shown in the paper, the registration precision can reach the 1~2 pixels level of the optical images.

Utterance Level Feature Aggregation with Deep Metric Learning for Speech Emotion Recognition.

Emotion is a form of high-level paralinguistic information that is intrinsically conveyed by human speech. Automatic speech emotion recognition is an essential challenge for various applications; including mental disease diagnosis; audio surveillance; human behavior understanding; e-learning and human–machine/robot interaction. In this paper, we introduce a novel speech emotion recognition method, based on the Squeeze and Excitation ResNet (SE-ResNet) model and fed with spectrogram inputs. In order to overcome the limitations of the state-of-the-art techniques, which fail in providing a robust feature representation at the utterance level, the CNN architecture is extended with a trainable discriminative GhostVLAD clustering layer that aggregates the audio features into compact, single-utterance vector representation. In addition, an end-to-end neural embedding approach is introduced, based on an emotionally constrained triplet loss function. The loss function integrates the relations between the various emotional patterns and thus improves the latent space data representation. The proposed methodology achieves 83.35% and 64.92% global accuracy rates on the RAVDESS and CREMA-D publicly available datasets, respectively. When compared with the results provided by human observers, the gains in global accuracy scores are superior to 24%. Finally, the objective comparative evaluation with state-of-the-art techniques demonstrates accuracy gains of more than 3%.

Noisy-LSTM: Improving Temporal Awareness for Video Semantic Segmentation

Semantic video segmentation is a key challenge for various applications. This paper presents a new model named Noisy-LSTM, which is trainable in an end-to-end manner, with convolutional LSTMs (ConvLSTMs) to leverage the temporal coherence in video frames, together with a simple yet effective training strategy that replaces a frame in a given video sequence with noises. Our training strategy spoils the temporal coherence in video frames and thus makes the temporal links in ConvLSTMs unreliable; this may consequently improve the ability of the model to extract features from video frames and serve as a regularizer to avoid overfitting, without requiring extra data annotations or computational costs. Experimental results demonstrate that the proposed model can achieve state-of-the-art performances on both the CityScapes and EndoVis2018 datasets. The code for the proposed method is available at https://github.com/wbw520/NoisyLSTM.

Nondestructive growing nano-ZnO on aramid fibers to improve UV resistance and enhance interfacial strength in composites

It is a great challenge in various applications to improve the resistances of aramid fibers (AFs) to harsh environments and enhance the interfacial bonding strength in composites along with maintaining the excellent mechanical properties. Herein, a simple nondestructive method was developed to prepare modified AF with zinc oxide (ZnO) micro-nano structures. Assisted by the adhesion and secondary functionalization of poly L-3,4-dihydroxyphenylalanine (PDOPA), ZnO nanoparticles (NPs) were successfully “seeded” onto the AF surface, and subsequently ZnO nanowires (NWs) were grown on the seed layer. As a result, the surface modification did not cause the significant changes of the tensile strength, and the surface roughness and UV resistance of the modified fibers were significantly improved. After 168 h of UV exposure, the retention of the tensile strength for the modified fiber was as high as 97.2%. Comparing with the pristine AF-reinforced composite, the interfacial shear strengths of AF-PDOPA-ZnO NP/epoxy resin and AF-PDOPA-ZnO NW/epoxy resin composites increased by 31.8% and 37.8%, respectively. Therefore, this effectively nondestructive modification method to grow ZnO nanostructures on a polymer substrate exhibits an extensive application prospect for organic fibers and their composites.

A Hybrid Generative Model for Online User Behavior Prediction

With the increase of rich datasets from various online platforms, predicting user behavior has been one of the most active research topics. The user behavior on these online platforms includes listening to music, watching videos, purchasing products, checking-in to places, and joining online sub-communities. Predicting online user behavior is an important challenge for various applications. Personalization, recommendation systems, target advertisements are based on user behavior prediction, where user's next purchases or actions need to be predicted. In this paper, we propose a hybrid generative model that can predict user behavior considering multiple factors. While previous work has been focused on two aspects individually: predicting repeat behavior or predicting new behavior, our model considers both aspects simultaneously during the learning process. The user-specific preference component is used to capture repeat behavior patterns, while the latent group preference component is used to discover new behavior. Besides these two components, we also consider the exogenous effect, which is not captured in the former two. Our experimental results on real-world datasets show how our proposed model outperforms the state-of-the-art model.

Controllable size and form of droplets in microfluidic-assisted devices: Effects of channel geometry and fluid velocity on droplet size.

Droplet-based microfluidic assisted devices have proposed an extensive interest in many applications such as lab-on-a-chip technologies as well as chemical/biological/nanomaterial preparation, chemical engineering, drug delivery, tissue engineering and biosensing. Here, a computational fluid dynamic model was developed for deep understanding of the droplet size and formation in a flow-focusing (FF) microchannel with consideration of the continuous phase (non-Newtonian fluid). The simulations presented an alternative method to achieve insights into this complicated process. In the following for the first time, the role of channel geometry, channel aspect ratio and flow rate ratio on droplet features including the mechanism of droplet formation, diameter/volume of droplet, velocity/amount of droplet formation, and final shape/size of the generated droplets were fully described. These findings could remarkably derive desirable protocols to control droplets characteristics comprising their size and shape in non-Newtonian fluids. Moreover, level set (LS) method was used for scrutinizing the droplet-breaking procedure in the microfluidic FF devices. The results showed that different droplet sizes could be prepared with changing the various parameters, demonstrating many challenges in various applications including lab-on-a-chip, cell encapsulation, drug delivery, tissue engineering, biosensing and bioimaging could be successfully addressed.

Choice of Application Layer Protocols for Next Generation Video Surveillance Using Internet of Video Things

Video surveillance has become ubiquitous due to the increasing security requirements in every sphere of life. The next generation video surveillance system (VSS) possesses great challenges in various applications, such as intelligent urban surveillance systems and smart cities. In these applications, we need to deal with the fast-growing number of surveillance nodes which introduce several constraints, e.g., high latency, high bandwidth, high energy consumption, and CPU and memory usage. To address these issues, the Internet of Video Things (IoVT), which is considered to be a part of the Internet of Things (IoT), can be a solution. The IoVT is composed of visual sensors (i.e., cameras) connected to the Internet. Unlike conventional systems, the VSS under an IoVT framework provides multiple layers (i.e., edge, fog, and cloud) of communication and decision making by capturing and analyzing rich contextual and behavioral information. Since an appropriate application layer protocol (ALP) can help in alleviating the challenges of future VSSs, the selection of ALPs is important for IoVT-based systems. Therefore, this paper presents a generic architecture of an IoVT-based VSS and a comparative analysis of several ALPs, such as MQTT, AMQP, HTTP, XMPP, CoAP, and DDS, with real-time experimentation. This analysis will assist the users to choose the appropriate ALPs in various surveillance applications and determine their suitability at different nodes of the IoVT framework.

A robust fuzzy least squares twin support vector machine for class imbalance learning

Twin support vector machine is one of the most prominent techniques for classification problems. It has been applied in various real world applications due to its less computational complexity. In most of the applications on classification, there is imbalance in the number of samples of the classes which leads to incorrect classification of the data points of the minority class. Further, while dealing with imbalanced data, noise poses a major challenge in various applications. To resolve these problems, in this paper we propose a robust fuzzy least squares twin support vector machine for class imbalance learning termed as RFLSTSVM-CIL using 2-norm of the slack variables which makes the optimization problem strongly convex. In order to reduce the effect of outliers, we propose a novel fuzzy membership function specifically for class imbalance problems. Our proposed function gives the appropriate weights to the datasets and also incorporates the knowledge about the imbalance ratio of the data. In our proposed model, a pair of system of linear equations is solved instead of solving a quadratic programming problem (QPP) which makes our model efficient in terms of computation complexity. To check the performance of our proposed approach, several numerical experiments are performed on synthetic and real world benchmark datasets. Our proposed model RFLSTSVM-CIL has shown better generalization performance in comparison to the existing methods in terms of AUC and training time.

Robust Composite High-Order Super-Twisting Sliding Mode Control of Robot Manipulators

This paper describes the design of a robust composite high-order super-twisting sliding mode controller (HOSTSMC) for robot manipulators. Robot manipulators are extensively used in industrial manufacturing for many complex and specialized applications. These applications require robots with nonlinear mechanical architectures, resulting in multiple control challenges in various applications. To address this issue, this paper focuses on designing a robust composite high-order super-twisting sliding mode controller by combining a higher-order super-twisting sliding mode controller as the main controller with a super-twisting higher-order sliding mode observer as unknown state measurement and uncertainty estimator in the presence of uncertainty. The proposed method adaptively improves the traditional sliding mode controller (TSMC) and the estimated state sliding mode controller (ESMC) to attenuate the chattering. The effectiveness of a HOSTSMC is tested over six degrees of freedom (DOF) using a Programmable Universal Manipulation Arm (PUMA) robot manipulator. The proposed method outperforms the TSMC and ESMC, yielding 4.9% and 2% average performance improvements in the output position root-mean-square (RMS) error and average error, respectively.

A Collaborative and Predictive Localization Algorithm for Wireless Sensor Networks

Accurate locating for the mobile target remains a challenge in various applications of wireless sensor networks (WSNs). Unfortunately, most of the typical localization algorithms perform well only in the WSN with densely distributed sensor nodes. The non-localizable problem is prone to happening when a target moves into the WSN with sparsely distributed sensor nodes. To solve this problem, we propose a collaborative and predictive localization algorithm (CPLA). The Gaussian mixture model (GMM) is introduced to predict the posterior trajectory for a mobile target by training its prior trajectory. In addition, the collaborative and predictive schemes are designed to solve the non-localizable problems in the two-anchor nodes locating, one-anchor node locating and non-anchor node locating situations. Simulation results prove that the CPLA exhibits higher localization accuracy than other tested predictive localization algorithms either in the WSN with sparsely distributed sensor nodes or in the WSN with densely distributed sensor nodes.

Ultrasonic Spray Deposition of Metal Oxide Films on High Aspect Ratio Microstructures for Three-Dimensional All-Solid-State Li-ion Batteries

Deposition of functional materials on nonplanar surfaces remains a challenge for various applications, including three-dimensional (3D) all-solid-state Li-ion batteries. In this Letter we present a new process to deposit functional oxide materials on high aspect ratio microstructures without the use of vacuum-based deposition methods. Using ultrasonic spray deposition in combination with metal citrate chemistry, we were able to deposit high-quality coatings on Si microcylinders with an aspect ratio of 10. These results were achieved by controlling the precursor chemistry, wetting properties, gel mobility, and precursor decomposition. The versatility of the process was shown by depositing titanium oxide (TiO2), lithium lanthanum titanate (Li0.35La0.55TiO3), and tungsten oxide (WO3) on Si microcylinders of 50 μm length with an intercylinder distance of 5 μm. Finally, a proof of the 3D battery concept was achieved by coating of TiN/Si microcylinders with WO3 using a minimized thermal budget to preserve the (...

An Improved End to End System for Sharing and Managing Data in Cloud Storage

Cloud Computing is one of the emerging area in modern times. It is providing excellent facilities due to its adaptable infrastructure and features, but the issues related to trust management and security are key challenges in various applications. Whenever data or information is shared or transmitted from one host to another in cloud storage there is a need for security and trust assurances to prevent data from unauthorized hosts. In this article we present a survey which addresses the various challenges of cloud storage along with the different trust computation models to identify important research issues in the hotfoot growing area of computer technology. In order to resolve those issues the proposed methodology incorporates the trust evaluation and Digital Envelope techniques for preserving data against the un-trusted hosts and networks.

From nano-cubic particle to micro-spindle aggregation: The control of long chain fatty acid on the morphology of calcium carbonate

The morphological control of calcium carbonate (CaCO3) colloidal particles, which allows us to obtain specific material properties, is one of the key challenges in various applications. To this end, we have studied the formation of CaCO3 particles by adding different long chain fatty acids—lauric acid, palmitic acid and stearic acid—into calcium hydroxide (Ca(OH)2) slurry, and then, by dispersing carbon dioxide (CO2) gas into the slurry via a micropore plate. Their polymorph, morphology and surface feature were characterized by XRD, TEM and FTIR, respectively. It is shown that the crystalline phase of CaCO3 particles leaves unchanged with respect to the used species and amounts of fatty acids, what is changed is the shape of CaCO3 particles from nano-cubic to micro-spindle. Thus we underlined the tuning mechanism of precipitation.

Tracking Amino Acids in Chiral Quantum Corrals

Molecular diffusion, motion, and conformation are critical to chemical and biological processes. Concurrently, understanding how chirality affects these processes has become a critical challenge for various applications in the pharmaceutical and food industries ranging from drug catalysis to novel sensing. Here, we present a unique way of transferring the chirality of simple amino acids, l- and d-alanine, to large-scale chiral networks on Cu(111). We further utilize the unique geometry of the chiral network as a scaffolding to isolate individual molecules within a 1.2 nm hexagonal pore. These hexagonal pores act as single molecule “race tracks” where excess alanine molecules trapped at the perimeter are observed to hop between six distinct locations around the perimeter. Scanning tunneling microscopy (STM) as well as density functional theory (DFT) calculations have been utilized to directly track, influence, and probe this molecular motion confined to self-assembled, chiral, hexagonal pores which also form quantum corrals.

Effective UV surface albedo of seasonally snow-covered lands

Abstract. At ultraviolet wavelengths the albedo of most natural surfaces is small with the striking exception of snow and ice. Therefore, snow cover is a major challenge for various applications based on radiative transfer modelling. The aim of this work was to determine the characteristic effective UV range surface albedo of various land cover types when covered by snow. First we selected 1 by 1 degree sample regions that met three criteria: the sample region contained dominantly subpixels of only one land cover type according to the 8 km global land cover classification product from the University of Maryland; the average slope of the sample region was less than 2 degrees according to the USGS's HYDRO1K slope data; the sample region had snow cover in March according to the NSIDC Northern Hemisphere weekly snow cover data. Next we generated 1 by 1 degree gridded 360 nm surface albedo data from the Nimbus-7 TOMS Lambertian equivalent reflectivity data, and used them to construct characteristic effective surface albedo distributions for each land cover type. The resulting distributions showed that each land cover type experiences a characteristic range of surface albedo values when covered by snow. The result is explained by the vegetation that extends upward beyond the snow cover and masks the bright snow covered surface.