Search

Sound event detection (SED) and acoustic scene classification (ASC) are two key tasks related to each other in the field of computational auditory scene analysis. For example, during sound event detection, scene information can be used to exclude sound events that are unlikely to occur in this scene. In other words, scene information can improve the accuracy of sound event detection. However, existing works rarely detect sound events by considering acoustic scene information. Based on the internal relationship between sound events and scene information, this paper proposes a scene-dependent sound event detection (SDSED) approach, which combines scene information and sound event information using multi-task learning. In the proposed approach, we share common feature representation for the two tasks simultaneously. Meanwhile, a temporal attention mechanism is used to extract informative features from sound recordings. We test the proposed approach on Synthetic Sound Scenes dataset. Experimental results show that our proposed approach outperforms the state-of-the-art approaches. Compared with the referenced approach, our approach improves the segment-based F-score by 4.29% and reduces the segment-based error rate by 4.8%.

The 2019 novel coronavirus(COVID-19) spreads rapidly, and the large-scale infection leads to the lack of medical resources. For the purpose of providing more reasonable medical service to COVID-19 patients, we designed an novel adjuvant therapy system integrating warning, therapy, and post-therapy psychological intervention. The system combines data analysis, communication networks and artificial intelligence(AI) to design a guidance framework for the treatment of COVID-19 patients. Specifically, in this system, we first can use blood characteristic data to help make a definite diagnosis and classify the patients. Then, the classification results, together with the blood characteristics and underlying diseases disease characteristics of the patient, can be used to assist the doctor in treat treating the patient according to AI algorithms. Moreover, after the patient is discharged from the hospital, the system can monitor the psychological and physiological state at the data collection layer. And in the data feedback layer, this system can analyze the data and report the abnormalities of the patient to the doctor through communication network. Experiments show the effectiveness of our proposed system.

Visual scenes containing text in the TextVQA task require a simultaneous understanding of images, questions, and text in images to reason answers. However, most existing cross-modal tasks merely involve two modalities. There are thus few methods for modeling interactions across three modalities. To bridge this gap, we propose in this work cross- and intra-modal interaction modules for multiple (more than two) modalities, where scaled dot-product attention method is applied to model inter- and intra-modal relationship. In addition, we introduce guidance information to assist the attention method to learn a more accurate relationship distribution. We construct a Multi-level Complete Interaction (MLCI) model for the TextVQA task via stacking multiple blocks composed of our proposed interaction modules. We design a multi-level feature joint prediction approach to exploit output representations from each block in a complementary way to predict answers. The experimental results on the TextVQA dataset show that our model obtains a 5.42% improvement in accuracy more than the baseline. Extensive ablation studies are carried out for the comprehensive analysis of the proposed method. Our code is publicly available at https://github.com/zhangshengHust/mlci .

VQA attracts lots of researchers in recent years. It could be potentially applied to the remote consultation of COVID-19. Attention mechanisms provide an effective way of utilizing visual and question information selectively in visual question and answering (VQA). The attention methods of existing VQA models generally focus on spatial dimension. In other words, the attention is modeled as spatial probabilities that re-weights the image region or word token features. However, feature-wise attention cannot be ignored, as image and question representations are organized in both spatial and feature-wise modes. Taking the question “What is the color of the woman’s hair” for example, identifying the hair color attribute feature is as important as focusing on the hair region. In this paper, we propose a novel neural network module named “multimodal feature-wise attention module” (MulFA) to model the feature-wise attention. Extensive experiments show that MulFA is capable of filtering representations for feature refinement and leads to improved performance. By introducing MulFA modules, we construct an effective union feature-wise and spatial co-attention network (UFSCAN) model for VQA. Our evaluation on two large-scale VQA datasets, VQA 1.0 and VQA 2.0, shows that UFSCAN achieves performance competitive with state-of-the-art models.

Currently, many deep learning (DL)-based low-dose CT image postprocessing technologies fail to consider the anatomical differences in training data among different human body sites, such as the cranium, lung and pelvis. In addition, we can observe evident anatomical similarities at the same site among individuals. However, these anatomical differences and similarities are ignored in the current DL-based methods during the network training process. In this paper, we propose a deep network trained by introducing anatomical site labels, termed attributes for training data. Then, the network can adaptively learn to obtain the optimal weight for each anatomical site. By doing so, the proposed network can take full advantage of anatomical prior information to estimate high-resolution CT images. Furthermore, we employ a Wasserstein generative adversarial network (WGAN) augmented with attributes to preserve more structural details. Compared with the traditional networks that do not consider the anatomical prior and whose weights are consequently the same for each anatomical site, the proposed network achieves better performance by adaptively adjusting to the anatomical prior information.

To evaluate the characteristics of traffic noise under different network structures, an evaluation method is established to clarify the mechanism of traffic noise impacted by variational network structures. First, a stochastic user equilibrium (SUE) model considering intersection delays is developed to allocate traffic flows of different network structures, and a flow-based univariate noise prediction model is used to simulate network noise distribution. Secondly, differentiation indexes including the coefficient of inhomogeneity (COI) and the concept of Lane Area Ratio (LAR) are set to quantify the network structure differences. Finally, a structural equation model (SEM) is developed to investigate the influence mechanism of network structure differentiation on traffic noise. The following conclusions are obtained: (1) The impact of network density differentiation on traffic noise is mainly reflected in the changes of road traffic flow and speed. Traffic noise decreases as the network density increases. In this case, when network density increases by 1 km/km2, traffic noise decreases by 1.6 dB. As network density increases, which means a dispensation of traffic flow, traffic noise is diminished by a reduction in traffic flow and speed. (2) The impact of road spatial location differentiation on traffic noise mainly depends on the number of noise sources. Traffic noise increases with the dense distribution of roads. In this case, when the COI increases by 1, the traffic noise increases by 3.0 dB. A higher COI indicates that the region will be exposed to more noise sources, which leads to traffic noise raises. The study can provide an effective basis for traffic noise control at the initial stage of network planning.

Face hallucination (FH) aims to reconstruct high-resolution faces from low-resolution face inputs, making it significant to other face-related tasks. Different from general super resolution issue, it often requires facial priors other than general extracted features thus leading to fusion of more than one kind of feature. The existing CNN-based FH methods often fuse different features indiscriminately which may introduce noises. Also the latent relations among different features which may be useful are taken into less consideration. To address the above issues, we propose an end-to-end deep ensemble network which aggregates three extraction sub-nets in attention-based manner. In our ensemble strategy, both relations among different features and inter-dependencies among different channels are dug out through the exploitation of spatial attention and channel attention. And for the diversity of extracted features, we aggregate three different sub-nets, which are the basic sub-net for basic features, the auto-encoder sub-net for facial shape priors and the dense residual attention sub-net for fine-grained texture features. Conducted ablation studies and experimental results show that our method achieves effectiveness not only in PSNR (Peak Signal to Noise Ratio) and SSIM (Structural Similarity Index) metrics but more importantly in clearer details within both key facial areas and whole range. Also results show that our method achieves real-time hallucinating faces by generating one image in 0.0237s.

A light delivery system based on localized surface plasmon (LSP) is the key component of heat-assisted magnetic recording (HAMR). We demonstrated a novel resonator–antenna system as the light delivery system for HAMR. We numerically analyzed the interaction mechanism between the resonator waveguide and the antenna. The localized enhanced resonated field of the resonator can improve the electric-field enhancement of the antenna. Based on the interaction mechanism, we proposed the ring resonator to achieve higher electric-field enhancement of light spots as the resonated field can be several times higher than input light under the resonance condition. The input laser energy can be recycled by limiting the energy in the ring waveguide without using higher power laser sources. Numerical results show that the nanofocusing performance in the ring resonator system would be much better than normal light systems. The maximum electric-field enhancement of light spots can be 124, and the absorption efficiency can be about 10% in the ring-lollipop system. The proposed resonator–antenna system can effectively increase the nanofocusing performance of the light delivery system for HAMR.

We have successfully fabricated Pt/BiFeO 3 / SrRuO 3 /SrTiO 3 resistive devices with magnetron sputtering at different deposition temperatures and thicknesses. The resistive behaviors and conduction mechanisms can be modulated by the thickness and deposition temperatures of BFO films. BFO films with smaller thickness and higher deposition temperature show larger memory window and lower switching voltage possibly due to the more oxygen vacancies and defects.

Various schemes have been investigated to study the modeling of recording media, the writing/reading process, and the information recovery in magnetic recording systems, such as two-dimensional magnetic recording (TDMR) and bit-patterned media recording (BPMR). As the bit scales down to the size of the media particles, the channel modeling and simulation become more complicated because of the increasing media noise. Meanwhile, it relies on more powerful 2-D signal-processing algorithms error control codes. The contributions of this article are three-fold: 1) we improve the read process in terms of our proposed reliability for read response that captures the effects of inter-track interference (ITI), inter-symbol interference (ISI), and the randomness of grains; 2) we present a neural network block predictor (NNBP) and a logistic regression method to provide heuristic methods in studying data recovery; and 3) we investigate the most possible harmful transitions and patterns. Results show the NNBP achieves good performance and is good at detecting those totally isolated bits that are thought to be bad patterns from an intuitive point of view.

Rapid growth of IoT applications and their interference in our daily lives led to many different IoT devices which generates enormous data. The IoT devices’ resources are very limited, so storing and processing IoT data in the devices is very inefficient. Several resources of cloud-computing are efficiently used to handle some IoT resources issues. While using resources in the cloud centers cause some other issues, like latency in the IoT applications, which are time-critical. Thus, the technology of edge cloud has evolved recently. This technology permits storage and data processing at the network edge. This paper studies edge computing in-depth for timeless sensitive devices in IoT. In-depth, cutting-edge IoT computing systems (ECAs-IoT) are evaluated and characterized in this paper according to numerous criteria, such as information placement, improvisation facilities, reliability, and data visualization. Moreover, according to distinctive properties, the paper aims at comparing each structure in detail. The paper also highlights the significant limitations of the new ECAs-IoT and recommends solutions to them. The studies also introduce and propose solutions to some of the most important restrictions of the current ECAs-IoT. Consequently, in the edge computing domain, this survey outlines the IoT implementations. Lastly, with the use of IoT implementations for ECAs-IoT, the paper suggests four distinct scenarios.

Two-dimensional magnetic recording (TDMR) by shingled magnetic recording (SMR) is one of the promising technologies to achieve a magnetic recording areal density (AD) of 10 Tb/in2. The readback signal of TDMR suffers severe 2-D inter-symbol interference (2-D ISI). At ultra-high AD, reader array is proposed to process the signals from the 2-D ISI channel because it can acquire multiple interfered signal streams simultaneously from the adjacent tracks. In this article, we propose a novel architecture for signal detection, which uses a four-reader array to recover the recording data. The auxiliary reader of the four-reader array can offer additional soft information of bits on the detected track, and this information can be used to improve the bit error rate (BER) performance. We consider three approaches to utilize the auxiliary reader and investigate the position of the auxiliary reader. Simulation results show that the proposed approach outperforms both the conventional single-reader detection and three-reader array detection. The auxiliary reader is set with an offset of 10% track pitch and can obtain a significant improvement of BER performance when the signal-to-noise rates (SNRs) of all the four readers are the same.