Ansari, M, Jones, B, Zhu, H, Shariati, N & Guo, YJ 2021, 'A Highly Efficient Spherical Luneburg Lens for Low Microwave Frequencies Realized with a Metal-based Artificial Medium', IEEE Transactions on Antennas and Propagation.
View/Download from: Publisher's site
View description>>
IEEE This paper describes a novel spherical lens antenna constructed of planar layers of light-weight foam with equally spaced conducting inclusions of varying sizes on an orthogonal grid. This construction largely overcomes the problems of weight and cost that have tended to make larger low frequency Luneburg lenses impractical. A penalty for this type of design is that some anisotropy exists in the lens’s dielectric. This effect is examined using both ray tracing techniques and full-wave simulation and it is found that the principal consequence is that the focal length of the lens varies in different directions. Methods for mitigating the effect are proposed. A prototype lens antenna intended for cellular use in the band 3.3 – 3.8 GHz with dual linear slant polarized feeds was designed and constructed to confirm the findings. Measured results show a peak gain of 23 dBi which is less than 1 dB lower than the maximum possible directivity from the lens’s cross section area. Scanning loss is less than 0.8 dB over the whole sphere. Simulated and measured performance show excellent agreement over the whole sphere. The overall performance of the prototype lens antenna demonstrates that this type of lens should be very suitable for use in high-gain multibeam antennas at lower microwave frequencies.
Chakraborty, S, Milner, LE, Zhu, X, Sevimli, O, Parker, AE & Heimlich, MC 2021, 'An Edge-Coupled Marchand Balun with Partial Ground for Excellent Balance in 0.13 μm SiGe Technology', IEEE Transactions on Circuits and Systems II: Express Briefs, vol. 68, no. 1, pp. 226-230.
View/Download from: Publisher's site
View description>>
© 2004-2012 IEEE. An edge-coupled meandered three-coupled-line Marchand balun with a partial ground plane implemented in 0.13 μ {m} SiGe Bi-CMOS technology is presented in this brief. The balance performance of the designed balun is significantly improved by creating a 'no ground plane' beneath the coupled-line structure, which is demonstrated by simulating two baluns: one with a partial ground and the other with a solid ground underneath. The measured amplitude and phase imbalances are less than 0.4 dB and 2.5°, across the 3-dB bandwidth from 21.5 to 95 GHz, surpassing previously reported results of edge-coupled Marchand baluns. The balun occupies 230 μ \text{m}\,\,×370\,\,μ {m}.
Chen, L, Liu, H, Hora, J, Zhang, JA, Yeo, KS & Zhu, X 2021, 'A Monolithically Integrated Single-Input Load-Modulated Balanced Amplifier With Enhanced Efficiency at Power Back-Off', IEEE Journal of Solid-State Circuits.
View/Download from: Publisher's site
View description>>
In this article, the design of a power amplifier (PA) using a simple but effective architecture, namely, load-modulated balanced amplifier (LMBA), is presented. Using this architecture for PA design, it can achieve not only a relatively high saturated output power but also an excellent efficiency enhancement at the power back-off (PBO) region. To prove that the presented approach is feasible in practice, a PA is designed in a 1-μm gallium arsenide (GaAs) HBT process. Operating under a 5-V power supply, the PA can deliver more than 31-dBm saturated output power with 36% collector efficiency (CE) at 5 GHz. Moreover, it also achieves 1.2 and 1.23 times CE enhancement over an idealistic Class-B PA at 6- and 9-dB PBO levels, respectively. Finally, the designed PA supports 64-quadrature amplitude modulation (QAM) with 80 Msys/s at 22-dBm average output power while still maintaining an error vector magnitude (EVM) and adjacent channel power ratio (ACPR) better than -29.5 dB and -29.4 dBc, respectively.
Chen, L, Liu, Y, Yang, S & Guo, YJ 2021, 'Efficient Synthesis of Filter-and-Sum Array With Scanned Wideband Frequency-Invariant Beam Pattern and Space-Frequency Notching', IEEE Signal Processing Letters.
View/Download from: Publisher's site
View description>>
IEEE This work generalizes the Fourier transform (FT)-based frequency-invariant beamforming (FIB) method to the synthesis of scanned frequency-invariant (FI) beam pattern with space-frequency notching for an array with non-isotropic elements. Wideband FI pattern characteristics are described by using multiple reference sub-band FI patterns that are obtained through an iterative single-frequency FT-based synthesis method. By applying fast Fourier transform (FFT) on the combination of these multiple reference sub-band FI patterns, a wideband excitation distribution can be generated. Based on this excitation distribution, we construct a new wideband excitation distribution that is conjugate-symmetric about zero frequency, so that real-valued finite-impulse-response (FIR) filter coefficients can be obtained by applying FFT on the constructed distribution. Two numerical examples are introduced to show the effectiveness and efficiency of the proposed method for synthesizing scanned FI beam patterns with complicated notching requirements.
Gao, G, Yu, Y, Xie, J, Yang, J, Yang, M & Zhang, J 2021, 'Constructing multilayer locality-constrained matrix regression framework for noise robust face super-resolution', PATTERN RECOGNITION, vol. 110.
View/Download from: Publisher's site
Ge, Z, Chen, L, Yang, L, Gomez-Garcia, R & Zhu, X 2021, 'On-Chip Millimeter-Wave Integrated Absorptive Bandstop Filter in (Bi)-CMOS Technology', IEEE ELECTRON DEVICE LETTERS, vol. 42, no. 1, pp. 114-117.
View/Download from: Publisher's site
Jin, D, Wen, S, Liu, Y, Wang, F, Lin, G, zhou, J, Shi, B & Suh, YD 2021, 'Nanobarcodes with multidimensional optical information beyond diffraction limit'.
View/Download from: Publisher's site
View description>>
Abstract Precise design and fabrication of heterogeneous nanostructures will enable nanoscale devices to integrate multiple desirable functionalities. But due to the diffraction limit (~200 nm), the optical uniformity and diversity within the heterogeneous functional nanostructures are hardly controlled and characterized. Here we report a set of nanobarcodes, each optically active section has its unique nonlinear responses to donut illumination patterns, so that one can discern each unit with super resolution. To achieve this, we first realized an approach of highly controlled epitaxial growth and produced a range of one-dimensional heterogeneous structures. Each section along the nanorod structure display tunable upconversion emissions, in four optically orthogonal dimensions, including colour, lifetime, excitation wavelength, and power dependency. Moreover, we demonstrated a 210 nm single nanorod as the smallest polychromatic light source for the on-demand generation of RGB photonic emissions. Remarkably, within a space of 50 nm, only 1/20th of the excitation wavelength, multiple codes can be successfully coded and decoded in 4 optical dimensions. This precision control enables the fabrication of super capacity geometrical barcodes with theoretical coding capacity up to (24-1)4. This work benchmarks our new ability towards the full control of sub-diffraction-limit optical diversities of single heterogeneous nanoparticles.
Liu, L, Jiang, J, Jia, W, Amirgholipour, S, Wang, Y, Zeibots, M & He, X 2021, 'DENet: A Universal Network for Counting Crowd With Varying Densities and Scales', IEEE Transactions on Multimedia, vol. 23, pp. 1060-1068.
View/Download from: Publisher's site
Ni, Z, Zhang, JA, Huang, X, Yang, K & Yuan, J 2021, 'Uplink Sensing in Perceptive Mobile Networks With Asynchronous Transceivers', IEEE TRANSACTIONS ON SIGNAL PROCESSING, vol. 69, pp. 1287-1300.
View/Download from: Publisher's site
Rahman, ML, Zhang, JA, Huang, X, Guo, YJ & Lu, Z 2021, 'Gaussian-Mixture-Model Based Clutter Suppression in Perceptive Mobile Networks', IEEE COMMUNICATIONS LETTERS, vol. 25, no. 1, pp. 152-156.
View/Download from: Publisher's site
Sang, L, Xu, M, Qian, S & Wu, X 2021, 'Knowledge graph enhanced neural collaborative recommendation', Expert Systems with Applications, vol. 164.
View/Download from: Publisher's site
View description>>
© 2020 Existing neural collaborative filtering (NCF) recommendation methods suffer from severe sparsity problem. Knowledge Graph (KG), which commonly consists of fruitful connected facts about items, presents an unprecedented opportunity to alleviate the sparsity problem. However, pure NCF models can hardly model the high-order connectivity in KG, and ignores complex pairwise correlations between user/item embedding dimensions. To address these problems, we propose a novel Knowledge graph enhanced Neural Collaborative Recommendation (K-NCR) framework, which effectively combines user–item interaction information and auxiliary knowledge information for recommendation task into three parts: (1) For items, the proposed propagating model learns the representation of item entity. It recursively aggregates information from its multi-hop neighbours in KG, and employs an attention mechanism to discriminate the importance of the relation type to mine users’ potential preferences. (2) For users, another heterogeneous attention weights are leveraged to strengthen the embedding learning of users. (3) The user and item embeddings are then fed into a newly designed two-dimensional interaction map with convolutional hidden layers to model the complex pairwise correlations between their embedding dimensions explicitly. Extensive experimental results on three benchmark datasets demonstrate the effectiveness of our K-NCR framework.
Saputra, YM, Hoang, DT, Nguyen, DN & Dutkiewicz, E 2021, 'A Novel Mobile Edge Network Architecture with Joint Caching-Delivering and Horizontal Cooperation', IEEE Transactions on Mobile Computing, vol. 20, no. 1, pp. 19-31.
View/Download from: Publisher's site
View description>>
Mobile edge caching/computing (MEC) has been emerging as a promising paradigm to provide ultra-high rate, ultra-reliable, and/or low-latency communications in future wireless networks. In this paper, we introduce a novel MEC network architecture that leverages the optimal joint caching-delivering with horizontal cooperation among mobile edge nodes (MENs). To that end, we first formulate the content-access delay minimization problem by jointly optimizing the content caching and delivering decisions under various network constraints (e.g., network topology, storage capacity and users’ demands at each MEN). However, the strongly mutual dependency between the decisions makes the problem a nested dual optimization that is proved to be NP-hard. To deal with it, we propose a novel transformation method to transform the nested dual problem to an equivalent mixed-integer nonlinear programming (MINLP) optimization problem. Then, we design a centralized solution using an improved branch-and-bound algorithm with the interior-point method to find the joint caching and delivering policy which is within 1 percent of the optimal solution. Since the centralized solution requires the full network topology and information from all MENs, to make our solution scalable, we develop a distributed algorithm which allows each MEN to make its own decisions based on its local observations. Extensive simulations demonstrate that the proposed solutions can reduce the total average delay for the whole network up to 40 percent compared with other current caching policies. Furthermore, the proposed solutions also increase the cache hit ratio for the network up to 4 times, thereby dramatically reducing the traffic load on the backhaul network
Shan, X, Wang, F, Wang, D, Wen, S, Chen, C, Di, X, Nie, P, Liao, J, Liu, Y, Ding, L, Reece, PJ & Jin, D 2021, 'Optical tweezers beyond refractive index mismatch using highly doped upconversion nanoparticles', Nature Nanotechnology.
View/Download from: Publisher's site
View description>>
© 2021, The Author(s), under exclusive licence to Springer Nature Limited. Optical tweezers are widely used in materials assembly1, characterization2, biomechanical force sensing3,4 and the in vivo manipulation of cells5 and organs6. The trapping force has primarily been generated through the refractive index mismatch between a trapped object and its surrounding medium. This poses a fundamental challenge for the optical trapping of low-refractive-index nanoscale objects, including nanoparticles and intracellular organelles. Here, we report a technology that employs a resonance effect to enhance the permittivity and polarizability of nanocrystals, leading to enhanced optical trapping forces by orders of magnitude. This effectively bypasses the requirement of refractive index mismatch at the nanoscale. We show that under resonance conditions, highly doping lanthanide ions in NaYF4 nanocrystals makes the real part of the Clausius–Mossotti factor approach its asymptotic limit, thereby achieving a maximum optical trap stiffness of 0.086 pN μm–1 mW–1 for 23.3-nm-radius low-refractive-index (1.46) nanoparticles, that is, more than 30 times stronger than the reported value for gold nanoparticles of the same size. Our results suggest a new potential of lanthanide doping for the optical control of the refractive index of nanomaterials, developing the optical force tag for the intracellular manipulation of organelles and integrating optical tweezers with temperature sensing and laser cooling7 capabilities.
Sun, Z, Yao, Y, Xiao, J, Zhang, L, Zhang, J & Tang, Z 2021, 'Exploiting textual queries for dynamically visual disambiguation', PATTERN RECOGNITION, vol. 110.
View/Download from: Publisher's site
Tang, Q, Yang, J, Liu, H, Guo, Z & Jia, W 2021, 'Single image deraining using Context Aggregation Recurrent Network', Journal of Visual Communication and Image Representation, vol. 75, pp. 103039-103039.
View/Download from: Publisher's site
Waheed, N, He, X, Ikram, M, Usman, M, Hashmi, SS & Usman, M 2021, 'Security and Privacy in IoT Using Machine Learning and Blockchain', ACM Computing Surveys, vol. 53, no. 6, pp. 1-37.
View/Download from: Publisher's site
View description>>
Security and privacy of users have become significant concerns due to the involvement of the Internet of Things (IoT) devices in numerous applications. Cyber threats are growing at an explosive pace making the existing security and privacy measures inadequate. Hence, everyone on the Internet is a product for hackers. Consequently, Machine Learning (ML) algorithms are used to produce accurate outputs from large complex databases, where the generated outputs can be used to predict and detect vulnerabilities in IoT-based systems. Furthermore, Blockchain (BC) techniques are becoming popular in modern IoT applications to solve security and privacy issues. Several studies have been conducted on either ML algorithms or BC techniques. However, these studies target either security or privacy issues using ML algorithms or BC techniques, thus posing a need for a combined survey on efforts made in recent years addressing both security and privacy issues using ML algorithms and BC techniques. In this article, we provide a summary of research efforts made in the past few years, from 2008 to 2019, addressing security and privacy issues using ML algorithms and BC techniques in the IoT domain. First, we discuss and categorize various security and privacy threats reported in the past 12 years in the IoT domain. We then classify the literature on security and privacy efforts based on ML algorithms and BC techniques in the IoT domain. Finally, we identify and illuminate several challenges and future research directions using ML algorithms and BC techniques to address security and privacy issues in the IoT domain.
Wang, N, Zhang, J-N, Liu, Z-Y, Ding, C, Sui, G-R, Jia, H-Z & Gao, X-M 2021, 'An Enhanced Thermoelectric Collaborative Cooling System With Thermoelectric Generator Serving as a Supplementary Power Source', IEEE Transactions on Electron Devices, pp. 1-8.
View/Download from: Publisher's site
Wang, X, Zha, X, Ni, W, Liu, RP, Guo, YJ, Niu, X & Zheng, K 2021, 'Game Theoretic Suppression of Forged Messages in Online Social Networks', IEEE Transactions on Systems, Man, and Cybernetics: Systems, pp. 1-11.
View/Download from: Publisher's site
Wei, Q, Feng, D & Jia, W 2021, 'UDR: An Approximate Unbiased Difference-Ratio Edge Detector for SAR Images', IEEE Transactions on Geoscience and Remote Sensing, pp. 1-18.
View/Download from: Publisher's site
Wong, S-W, Lin, J-Y, Yang, Y, Guo, Z-C, Zhu, L & Chu, Q-X 2021, 'Waveguide Components Based on Multiple-Mode Resonators', IEEE MICROWAVE MAGAZINE, vol. 22, no. 2, pp. 33-45.
View/Download from: Publisher's site
Wu, K, Zhang, JA, Huang, X, Guo, YJ & Heath, RW 2021, 'Waveform Design and Accurate Channel Estimation for Frequency-Hopping MIMO Radar-Based Communications', IEEE TRANSACTIONS ON COMMUNICATIONS, vol. 69, no. 2, pp. 1244-1258.
View/Download from: Publisher's site
Xi, Y, Jia, W, Zheng, J, Fan, X, Xie, Y, Ren, J & He, X 2021, 'DRL-GAN: Dual-Stream Representation Learning GAN for Low-Resolution Image Classification in UAV Applications', IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.
View/Download from: Publisher's site
View description>>
CCBY Identifying tiny objects from extremely low resolution (LR) UAV-based remote sensing images is generally considered as a very challenging task, because of very limited information in the object areas. In recent years, there have been very limited attempts to approach this problem. These attempts intend to deal with LR image classification by enhancing either the poor image quality or image representations.In this paper, we argue that the performance improvement in LR image classification is affected by the inconsistency of the information loss and learning priority on Low-Frequency (LF) components and High-Frequency (HF) components.To address this LF-HF inconsistency problem, we propose a Dual-Stream Representation Learning Generative Adversarial Network (DRL-GAN).The core idea is to produce super image representations optimal for LR recognition by simultaneously recovering the missing information in LF and HF components, respectively, under the guidance of high-resolution (HR) images.We evaluate the performance of DRL-GAN on the challenging task of LR image classification.A comparison of the experimental results on the LR benchmark, namely HRSC and CIFAR-10, and our newly collected “WIDER-SHIP” dataset demonstrates the effectiveness of our DRL-GAN, which significantly improves the classification performance, with up to 10% gain on average.
Xu, JX, Huang, M, Li, HY, Yang, Y & Zhang, XY 2021, 'Design of Balanced Filtering Rat-Race Coupler Based on Quad-Mode Dielectric Resonator', IEEE Transactions on Circuits and Systems II: Express Briefs.
View/Download from: Publisher's site
View description>>
A balanced filtering rat-race coupler is proposed based on a quad-mode dielectric resonator (DR) in this paper. Due to the electromagnetic-field distributions of the DR, the desired balanced performance with high common-mode suppression is realized without additional circuits. Moreover, by setting the feeding probes at the proper locations, the quad-mode DR is also used to obtain the desired coupler topology with filtering responses. Accordingly, a balanced filtering rat-race coupler, integrating three circuit functions of the balanced circuit, rat-race coupler, and bandpass filter, can be designed into a single-cavity configuration, resulting in a very compact size. For verification, the proposed circuit is implemented, which shows good filtering responses, low amplitude imbalance, high common-mode suppression, and excellent phase characteristics.
Yang, T, Ding, C, Ziolkowski, RW & Guo, YJ 2021, 'An Epsilon-Near-Zero (ENZ) Based, Ultra-Wide Bandwidth Terahertz Single-Polarization Single-Mode Photonic Crystal Fiber', JOURNAL OF LIGHTWAVE TECHNOLOGY, vol. 39, no. 1, pp. 223-232.
View/Download from: Publisher's site
Yao, L, Kusakunniran, W, Wu, Q, Zhang, J, Tang, Z & Yang, W 2021, 'Robust gait recognition using hybrid descriptors based on Skeleton Gait Energy Image', Pattern Recognition Letters.
View/Download from: Publisher's site
View description>>
© 2019 Gait features have been widely applied in human identification. The commonly-used representations for gait recognition can be roughly classified into two categories: model-free features and model-based features. However, due to the view variances and clothes changes, model-free features are sensitive to the appearance changes. For model-based features, there is great difficulty in extracting the underlying models from gait sequences. Based on the confidence maps and the part affinity fields produced by a two-branch multi-stage CNN network, a new model-based representation, Skeleton Gait Energy Image (SGEI), has been proposed in this paper. Another contribution is that a hybrid representation has been produced, which uses SGEI to remedy the deficiency of model-free features, Gait Energy Image (GEI) for instance. The experimental performances indicate that our proposed methods are more robust to the cloth changes, and contribute to increasing the robustness of gait recognition in the unconstrained environments with view variances and clothes changes.
You, X, Wang, CX, Huang, J, Gao, X, Zhang, Z, Wang, M, Huang, Y, Zhang, C, Jiang, Y, Wang, J, Zhu, M, Sheng, B, Wang, D, Pan, Z, Zhu, P, Yang, Y, Liu, Z, Zhang, P, Tao, X, Li, S, Chen, Z, Ma, X, Chih-Lin, I, Han, S, Li, K, Pan, C, Zheng, Z, Hanzo, L, Shen, XS, Guo, YJ, Ding, Z, Haas, H, Tong, W, Zhu, P, Yang, G, Wang, J, Larsson, EG, Ngo, HQ, Hong, W, Wang, H, Hou, D, Chen, J, Chen, Z, Hao, Z, Li, GY, Tafazolli, R, Gao, Y, Poor, HV, Fettweis, GP & Liang, YC 2021, 'Towards 6G wireless communication networks: vision, enabling technologies, and new paradigm shifts', Science China Information Sciences, vol. 64, no. 1.
View/Download from: Publisher's site
View description>>
© 2020, The Author(s). The fifth generation (5G) wireless communication networks are being deployed worldwide from 2020 and more capabilities are in the process of being standardized, such as mass connectivity, ultra-reliability, and guaranteed low latency. However, 5G will not meet all requirements of the future in 2030 and beyond, and sixth generation (6G) wireless communication networks are expected to provide global coverage, enhanced spectral/energy/cost efficiency, better intelligence level and security, etc. To meet these requirements, 6G networks will rely on new enabling technologies, i.e., air interface and transmission technologies and novel network architecture, such as waveform design, multiple access, channel coding schemes, multi-antenna technologies, network slicing, cell-free architecture, and cloud/fog/edge computing. Our vision on 6G is that it will have four new paradigm shifts. First, to satisfy the requirement of global coverage, 6G will not be limited to terrestrial communication networks, which will need to be complemented with non-terrestrial networks such as satellite and unmanned aerial vehicle (UAV) communication networks, thus achieving a space-air-ground-sea integrated communication network. Second, all spectra will be fully explored to further increase data rates and connection density, including the sub-6 GHz, millimeter wave (mmWave), terahertz (THz), and optical frequency bands. Third, facing the big datasets generated by the use of extremely heterogeneous networks, diverse communication scenarios, large numbers of antennas, wide bandwidths, and new service requirements, 6G networks will enable a new range of smart applications with the aid of artificial intelligence (AI) and big data technologies. Fourth, network security will have to be strengthened when developing 6G networks. This article provides a comprehensive survey of recent advances and future trends in these four aspects. Clearly, 6G with additional technical requireme...
Yu, G, Zhang, L, Wang, X, Yu, K, Ni, W, Zhang, JA & Liu, RP 2021, 'A novel Dual-Blockchained structure for contract-theoretic LoRa-based information systems', Information Processing and Management, vol. 58, no. 3.
View/Download from: Publisher's site
View description>>
LoRa serves as one of the most deployed technologies in Internet-of-Things-based information systems (IoT-IS), and self-motivated deployment is the key to the rollout of LoRa. Proper incentive can play an important role in encouraging the private deployment of LoRa, increasing coverage and promoting effective management of IoT-IS. However, existing incentive mechanisms have the vulnerabilities of insecure centralized architecture and excessive utility loss of LoRa Controllers and Gateways, due to asymmetric information between private owners of gateways and centralized controller (or service providers). Blockchain-based LoRa networks, as a promising solution, have not been comprehensively studied to address the vulnerabilities, let alone the other issues of security, scalability, and flexibility. In this paper, we propose a novel Dual-Chained LoRa-based information system (LoRa-IS) to provide globally cross-validated security. Behaviors, including state-of-the-art contract-theoretic incentive mechanism and new flow control protocol, can be secured with the tamper-resistance of Blockchains. Being part of the proposed incentive mechanism, the new self-driven flow control allows both the Dual-Chain system and the LoRa network to scale. To the best of our knowledge, the proposed system is the first comprehensive Blockchain-based LoRa-IS combined with contract theory. We also provide analysis and simulations, showing that our system can pay fair incentives under information asymmetry. With the new flow control, the system can optimize network coverage while improving the Blockchain scalability and flexibility.
Zhang, J, Liu, L, Wang, P & Zhang, J 2021, 'Exploring the Auxiliary Learning for Long-Tailed Visual Recognition', Neurocomputing.
View/Download from: Publisher's site
Zhang, YF, Zheng, J, Jia, W, Huang, W, Li, L, Liu, N, Li, F & He, X 2021, 'Deep RGB-D Saliency Detection without Depth', IEEE Transactions on Multimedia.
View/Download from: Publisher's site
View description>>
IEEE The existing saliency detection models based on RGB colors only leverage appearance cues to detect salient objects. Depth information also plays a very important role in visual saliency detection and can supply complementary cues for saliency detection. Although many RGB-D saliency models have been proposed, they require to acquire depth data, which is expensive and not easy to get. In this paper, we propose to estimate depth information from monocular RGB images and leverage the intermediate depth features to enhance the saliency detection performance in a deep neural network framework. Specically, we rst use an encoder network to extract common features from each RGB image and then build two decoder networks for depth estimation and saliency detection, respectively. The depth decoder features can be fused with the RGB saliency features to enhance their capability. Furthermore, we also propose a novel dense multiscale fusion model to densely fuse multiscale depth and RGB features based on the dense ASPP model. A new global context branch is also added to boost the multiscale features. Experimental results demonstrate that the added depth cues and the proposed fusion model can both improve the saliency detection performance. Finally, our model not only outperforms state-of-the-art RGB saliency models, but also achieves comparable results compared with state-of-the-art RGB-D saliency models.
Zhang, YF, Zheng, J, Li, L, Liu, N, Jia, W, Fan, X, Xu, C & He, X 2021, 'Rethinking feature aggregation for deep RGB-D salient object detection', Neurocomputing, vol. 423, pp. 463-473.
View/Download from: Publisher's site
View description>>
© 2020 Two-stream UNet based architectures are widely used in deep RGB-D salient object detection (SOD) models. However, UNet only adopts a top-down decoder network to progressively aggregate high-level features with low-level ones. In this paper, we propose to enrich feature aggregation via holistic aggregation paths and an extra bottom-up decoder network. The former aggregates multi-level features holistically to learn abundant feature interactions while the latter aggregates improved low-level features with high-level features, thus promoting their representation ability. Aiming at the two-stream architecture, we propose another early aggregation scheme to aggregate and propagate multi-modal encoder features at each level, thereby improving the encoder capability. We also propose a factorized attention module to efficiently modulate the feature aggregation action for each feature node with multiple learned attention factors. Experimental results demonstrate that all of the proposed components can gradually improve RGB-D SOD results. Consequently, our final SOD model performs favorably against other state-of-the-art methods.
Zheng, D, Zhang, H, Zhang, JA & Su, SW 2021, 'Stability and controller design of switched systems with sequence-based average dwell time', International Journal of Control.
View/Download from: Publisher's site
View description>>
© 2020 Informa UK Limited, trading as Taylor & Francis Group. This paper investigates the stability of both linear and nonlinear switched systems with sequence-based average preceding dwell time (SBAPDT) and sequence-based average subsequent dwell time (SBASDT). Different from existing work, we take the sequence of switch events into consideration. These two new analysis methods have the potential of reducing the conservativeness of the threshold time in continuous- and discrete-time systems. A comparative numerical example is also given to demonstrate the advantages of the new approaches. It shows that the designed controllers and calculated average dwell time thresholds are better than other methods.