Search

This paper presents a fused deep learning algorithm for ECG classification. It takes advantages of the combined convolutional and recurrent neural network for ECG classification, and the weight allocation capability of attention mechanism. The input ECG signals are firstly segmented and normalized, and then fed into the combined VGG and LSTM network for feature extraction and classification. An attention mechanism (SE block) is embedded into the core network for increasing the weight of important features. Two databases from different sources and devices are employed for performance validation, and the results well demonstrate the effectiveness and robustness of the proposed algorithm.

Currently, deep learning (DL) algorithms have achieved great success in many applications including computer vision and natural language processing. Many different kinds of DL models have been reported, such as DeepWalk, LINE, diffusionconvolutional neural networks, graph convolutional networks (GCN), and so on. The GCN algorithm is a variant of convolutional neural network and achieves significant superiority by using a one-order localised spectral graph filter. However, only a one-order polynomial in the Laplacian of GCN has been approximated and implemented, which ignores undirect neighbour structure information. The lack of rich structure information reduces the performance of the neural networks in the graph structure data. In this study, the authors deduce and simplify the formula of two-order spectral graph convolutions to preserve rich local information. Furthermore, they build a layerwise GCN based on this two-order approximation, i.e. two-order GCN (TGCN) for semi-supervised classification. With the two-order polynomial in the Laplacian, the proposed TGCN model can assimilate abundant localised structure information of graph data and then boosts the classification significantly. To evaluate the proposed solution, extensive experiments are conducted on several popular datasets including the Citeseer, Cora, and PubMed dataset. Experimental results demonstrate that the proposed TGCN outperforms the state-of-art methods.

Solution-based technology has potential in electrode patterning for low-cost, large-scale and flexible electronics. However, the formation of cracks often leads to unpredictable experimental results. This report explores the critical impact of solvent evaporation on the hollow crack formation and the adhesive property of silver nanoparticle film deposited using an ink-jet printable precursor ink. Uneven evaporation causes directional flows inside and along the surface of wet film, which contributes to nanoparticle film assembly initially at the periphery surface region. The area further away from the contact line tends to form thinner surface film, which can be easily deformed to form ripples under compressive stress. The suppressed evaporation of the liquid under the surface film makes the initially outward compensation flow reversed to inward, leaving a hollow inside morphology of the ripples, which can be transferred to cracks after complete solidification and will seriously degrade the adhesive property. This work has referential significance for the optimization of ink chemistry and film processing for high-quality film fabrication using solution-based method.

Inspired by the classic Sauvola local image thresholding approach, we systematically study it from the deep neural network (DNN) perspective and propose a new solution called SauvolaNet for degraded document binarization (DDB). It is composed of three explainable modules, namely, Multi-Window Sauvola (MWS), Pixelwise Window Attention (PWA), and Adaptive Sauolva Threshold (AST). The MWS module honestly reflects the classic Sauvola but with trainable parameters and multi-window settings. The PWA module estimates the preferred window sizes for each pixel location. The AST module further consolidates the outputs from MWS and PWA and predicts the final adaptive threshold for each pixel location. As a result, SauvolaNet becomes end-to-end trainable and significantly reduces the number of required network parameters to 40K -- it is only 1\% of MobileNetV2. In the meantime, it achieves the State-of-The-Art (SoTA) performance for the DDB task -- SauvolaNet is at least comparable to, if not better than, SoTA binarization solutions in our extensive studies on the 13 public document binarization datasets. Our source code is available at https://github.com/Leedeng/SauvolaNet., Comment: Submitted to 16th International Conference on Document Analysis and Recognition

With the rapid progress of deepfake techniques in recent years, facial video forgery can generate highly deceptive video contents and bring severe security threats. And detection of such forgery videos is much more urgent and challenging. Most existing detection methods treat the problem as a vanilla binary classification problem. In this paper, the problem is treated as a special fine-grained classification problem since the differences between fake and real faces are very subtle. It is observed that most existing face forgery methods left some common artifacts in the spatial domain and time domain, including generative defects in the spatial domain and inter-frame inconsistencies in the time domain. And a spatial-temporal model is proposed which has two components for capturing spatial and temporal forgery traces in global perspective respectively. The two components are designed using a novel long distance attention mechanism. The one component of the spatial domain is used to capture artifacts in a single frame, and the other component of the time domain is used to capture artifacts in consecutive frames. They generate attention maps in the form of patches. The attention method has a broader vision which contributes to better assembling global information and extracting local statistic information. Finally, the attention maps are used to guide the network to focus on pivotal parts of the face, just like other fine-grained classification methods. The experimental results on different public datasets demonstrate that the proposed method achieves the state-of-the-art performance, and the proposed long distance attention method can effectively capture pivotal parts for face forgery.

Inkjet printing silver electrodes for a-IGZO TFT array have recently become a hot spot benefitting from not requiring a conventional vacuum process. However, satellite droplets are unwanted in larg...

Marrying a randomly distributed network of wood fibers with sodium carboxymethyl cellulose (CMC-Na) by impregnation is a scalable method to prepare highly transparent and hazy (TH) paper with superior mechanical properties. However, the poor water resistance of this TH paper is a hindrance for its practical applications. In this study, a facile protonation process along with desalination is proposed to remove the above obstacle by reducing electrostatic repulsion among CMC molecular chains while strengthening the physical cross-linking among CMC molecular chains and also between CMC-H molecular chains and wood fibers in water. Consequently, the protonated TH paper not only presents an enhanced water resistance but also shows a negligible deterioration in optical and mechanical properties compared with the untreated TH paper. Moreover, the influence of the water resistance of various substrates (PET film, untreated TH paper, and protonated TH paper) on the electrical conductivity of inkjet-printed silver n...

We first verify the critical role of solvent evaporation on the resolution of inkjet printing. To confirm our hypothesis, we adjusted the evaporation rate gradient along the surface of adjacent droplets by controlling the drying microenvironment. Uneven solvent evaporation flux caused thermocapillary surface flow inward the space of micrometer-sized droplets and increase the air pressure, which prevented the neighboring droplets from coalescence. When reducing the droplet distance by the solvent evaporation-based method, a uniform profile could be obtained at the same time. This work brings us a step closer to resolving one of the critical bottlenecks to commercializing printed electronic goods.

Torrefaction is a promising and effective pretreatment technology to upgrade biomass into a more homogeneous and hydrophobic fuel with higher calorific value and energy density. The main objective of this study was to establish a torrefaction model to predict the basic properties (ultimate and proximate analysis, mass and energy yields) of torrefied biomass. Three groups of lignocellulosic biomass, namely woody biomass (Chinese fir), herbaceous biomass (corn stalk), and shell-like biomass (palm kernel shell), were used as raw material to establish the model. First, the simulation process was carried out on the entire torrefaction stage (200–300 °C) based on the linear fitting method. Then, in order to obtain a more accurate model, the model was established by using two-stage torrefaction process (mild stage with 200–250 °C and severe stage with 250–300 °C) and the quadratic equation. Further, the model was validated based on the torrefaction results from literature reports by using the analysis of variance. Result showed that the values of statistical parameters (relative error and mean absolute error) were less than 5%. This indicated that the simulated two-stage torrefaction model obtained by using the quadratic equation can, in general, aid in providing prediction of basic properties of woody torrefied biomass with higher accuracy. This study is helpful for the design of large scale torrefaction equipment and optimization of the torrefaction process.

Obtaining homogeneous surface profiles of inkjet-printed silver nanoparticle films (SNFs) is severely hampered by the classical coffee-ring phenomenon. Traditional approaches to suppress the coffee-ring effect are achieved by improving the uniformity of liquid evaporation rate or introducing inward Marangoni force during ink drying. However, these existing methods involve extra chemicals, treatments, or equipment that will definitely increase the production cost or reduce conductivity. In this study, we demonstrate an inexpensive and efficient method to obtain uniform surface profiles of inkjet-printed SNFs by rationally regulating the drying microenvironment. To this end, a number of surface profiles of printed dots and lines, such as concave, convex, and flat, were first obtained via this method. The underlying principle of this method was then investigated by analytical calculations based on the simplified vapor diffusion model. Consequently, homogeneous surface profiles of inkjet-printed SNFs were ach...

The control of channel length is of great significance in the fabrication of thin film transistors (TFTs) with high-speed operation. However, achieving short channel on untreated glass by traditional piezoelectric inkjet printing is problematic due to the impacting and rebounding behaviors of droplet impinging on solid surface. Here a novel method was proposed to obtain short channel length on untreated glass by taking advantage of the difference in the retraction velocities on both sides of an ink droplet. In addition, droplets contact mechanism was first introduced in our work to explain the formation of short channel in the printing process. Through printing droplets array with optimized drop space and adjusting appropriate printing parameters, a 2.4 μm of channel length for TFT, to the best of our knowledge, which is the shortest channel on substrate without pre-patterning, was achieved using piezoelectric inkjet printing. This study sheds light on the fabrication of short channel TFT for large size and high-resolution displays using inkjet printing technology.

Ultrashort channels of electrodes are essential for the construction of advanced functional devices with high-level integration and high operation speed. However, the channel length of fabricated electrodes is limited to 20 μm in inkjet printing. Although several methods have been previously proposed to obtain short channels, they require extra processing steps. In this paper, channel self-aligning phenomenon was observed in directly patterned electrodes on unmodified substrate by inkjet printing, when using an interspace defects growing method. Further exploring the underlying mechanism reveals that the capillary force induced air film prevents droplets coalescence, even on a substrate with no temperature differences. The wetting region, which is generated by the receding droplets impingement, will draw droplets closer together at a larger drop space, thus demanding smaller air pressure for coalescence inhibition and contributing to the self-aligning phenomenon of micro-sized droplets released by inkjet printing. Accordingly, an ultrashort channel of 2.38 μm is obtained with relatively smooth boundaries, when electrodes are printed on a slightly heated substrate, which reduces the air pressure between two neighboring droplets. This work will provide a significant reference for future high resolution applications of inkjet printing technology.

We first verify the critical impact of evaporation on the formation of zigzag hollow cracks by regulating the drying micro-environment of silver nanoparticle film. Uneven evaporation and component segregation contributes to the flows along the surface and inside of droplets. Asymmetric vapor concentration distribution is capable of weakening the surface flow of droplets, thus suppressing the inner compressive stress of nanoparticles and leading to a surface morphology with less cracks. Although defect-free and surface smooth nanoparticle film deposited by a solution-based method remains a big challenge, our work has referential significance to optimize high-quality nanoparticle film with appropriate deposition and curing processes. Moreover, an optimization possibility through the drying micro-environment should be considered in high-end applications due to its enhanced effect on high-resolution patterns. Electronic supplementary material The online version of this article (10.1186/s11671-018-2777-x) contains supplementary material, which is available to authorized users.

Representation-based classification methods such as sparse representation-based classification (SRC) and linear regression classification (LRC) have attracted a lot of attentions. In order to obtain the better representation, a novel method called projection representation-based classification (PRC) is proposed for image recognition in this paper. PRC is based on a new mathematical model. This model denotes that the 'ideal projection' of a sample point $x$ on the hyper-space $H$ may be gained by iteratively computing the projection of $x$ on a line of hyper-space $H$ with the proper strategy. Therefore, PRC is able to iteratively approximate the 'ideal representation' of each subject for classification. Moreover, the discriminant PRC (DPRC) is further proposed, which obtains the discriminant information by maximizing the ratio of the between-class reconstruction error over the within-class reconstruction error. Experimental results on five typical databases show that the proposed PRC and DPRC are effective and outperform other state-of-the-art methods on several vision recognition tasks., Comment: Accepted by the Thirty-Second AAAI Conference on Artificial Intelligence (AAAI-18)

It is of great importance to preserve locality and similarity information in semi-supervised learning (SSL) based applications. Graph based SSL and manifold regularization based SSL including Laplacian regularization (LapR) and Hypergraph Laplacian regularization (HLapR) are representative SSL methods and have achieved prominent performance by exploiting the relationship of sample distribution. However, it is still a great challenge to exactly explore and exploit the local structure of the data distribution. In this paper, we present an effect and effective approximation algorithm of Hypergraph p-Laplacian and then propose Hypergraph p-Laplacian regularization (HpLapR) to preserve the geometry of the probability distribution. In particular, p-Laplacian is a nonlinear generalization of the standard graph Laplacian and Hypergraph is a generalization of a standard graph. Therefore, the proposed HpLapR provides more potential to exploiting the local structure preserving. We apply HpLapR to logistic regression and conduct the implementations for remote sensing image recognition. We compare the proposed HpLapR to several popular manifold regularization based SSL methods including LapR, HLapR and HpLapR on UC-Merced dataset. The experimental results demonstrate the superiority of the proposed HpLapR., Comment: 9 pages, 6 figures

Recently, manifold regularized semi-supervised learning (MRSSL) received considerable attention because it successfully exploits the geometry of the intrinsic data probability distribution including both labeled and unlabeled samples to leverage the performance of a learning model. As a natural nonlinear generalization of graph Laplacian, p-Laplacian has been proved having the rich theoretical foundations to better preserve the local structure. However, it is difficult to determine the fitting graph p-Lapalcian i.e. the parameter which is a critical factor for the performance of graph p-Laplacian. Therefore, we develop an ensemble p-Laplacian regularization (EpLapR) to fully approximate the intrinsic manifold of the data distribution. EpLapR incorporates multiple graphs into a regularization term in order to sufficiently explore the complementation of graph p-Laplacian. Specifically, we construct a fused graph by introducing an optimization approach to assign suitable weights on different p-value graphs. And then, we conduct semi-supervised learning framework on the fused graph. Extensive experiments on UC-Merced data set demonstrate the effectiveness and efficiency of the proposed method., Comment: 13 pages, 7 figures. arXiv admin note: text overlap with arXiv:1806.08104

In this work, we investigated the effect of deposition conditions such as sputtering power, oxygen ratio and sputtering pressure on the properties of Nd–Al doped indium-zinc oxide (NA-IZO) thin-film transistors based on copper (Cu) electrodes. The quality of deposited films was evaluated by introducing a microwave photoconductivity decay (μ-PCD) measurement. The results showed strong connection between sputtering conditions and μ-PCD results, which reveal that sputtering process was vital for film quality control. In addition, the tendency of obtained Peak value and D value from μ-PCD was in highly consistent with the final devices performance especially the saturation mobility. As a result, the optimized devices showed a saturation mobility of 21.4 cm2 V−1 s−1, threshold voltage about 1.0 V, subthreshold swing about 0.18 V/decade and I on/I off ratio about 9.9 × 107. The non-destructive and efficient μ-PCD method offers a convenient way to optimize the deposition procedure for metal oxide semiconductors.

Printing technologies for thin-film transistors (TFTs) have recently attracted much interest owing to their eco-friendliness, direct patterning, low cost, and roll-to-roll manufacturing processes. Lower production costs could result if electrodes fabricated by vacuum processes could be replaced by inkjet printing. However, poor interfacial contacts and/or serious diffusion between the active layer and the silver electrodes are still problematic for achieving amorphous indium–gallium–zinc–oxide (a-IGZO) TFTs with good electrical performance. In this paper, silver (Ag) source/drain electrodes were directly inkjet-printed on an amorphous a-IGZO layer to fabricate TFTs that exhibited a mobility of 0.29 cm2·V−1·s−1 and an on/off current ratio of over 105. To the best of our knowledge, this is a major improvement for bottom-gate top-contact a-IGZO TFTs with directly printed silver electrodes on a substrate with no pretreatment. This study presents a promising alternative method of fabricating electrodes of a-IGZO TFTs with desirable device performance.

In this study, high performance amorphous In–Ga–Zn–O (a-IGZO) TFTs were successfully fabricated with inkjet-printed silver source-drain electrodes. The results showed that increased channel thickness has an improving trend in the properties of TFTs due to the decreased contact resistance. Compared with sputtered silver TFTs, devices with printed silver electrodes were more sensitive to the thickness of active layer. Furthermore, the devices with optimized active layer showed high performances with a maximum saturation mobility of 8.73 cm2 V−1 S−1 and an average saturation mobility of 6.97 cm2 V−1 S−1, I on/I off ratio more than 107 and subthreshold swing of 0.28 V/decade, which were comparable with the analogous devices with sputtered electrodes.

Recently, amorphous indium-gallium-zinc-oxide thin film transistors (a-IGZO TFTs) with inkjet printing silver source/drain electrodes have attracted great attention, especially for large area and flexible electronics applications. The silver ink could be divided into two types: one is based on silver nanoparticles, and the other is silver salt ink. Organic materials are essential in the formulation of nanoparticle ink as a strong disperse stabilizer to prevent agglomeration of silver particles, but will introduce contact problems between the silver electrodes and the a-IGZO active layer after annealing, which is difficult to eliminate and leads to poor device properties. Our experiment is aimed to reduce this effect by using a silver salt ink without stabilizer component. With optimized inkjet printing conditions, the high performance of a-IGZO TFT was obtained with a mobility of 4.28 cm2/V·s and an on/off current ratio over 106. The results have demonstrated a significant improvement for a-IGZO TFTs with directly printed silver electrodes. This work presents a promising platform for future printed electronic applications.

In this paper, we introduce a symmetric-key Latin square image cipher (LSIC) for grayscale and color images. Our contributions to the image encryption community include 1) we develop new Latin square image encryption primitives including Latin Square Whitening, Latin Square S-box and Latin Square P-box ; 2) we provide a new way of integrating probabilistic encryption in image encryption by embedding random noise in the least significant image bit-plane; and 3) we construct LSIC with these Latin square image encryption primitives all on one keyed Latin square in a new loom-like substitution-permutation network. Consequently, the proposed LSIC achieve many desired properties of a secure cipher including a large key space, high key sensitivities, uniformly distributed ciphertext, excellent confusion and diffusion properties, semantically secure, and robustness against channel noise. Theoretical analysis show that the LSIC has good resistance to many attack models including brute-force attacks, ciphertext-only attacks, known-plaintext attacks and chosen-plaintext attacks. Experimental analysis under extensive simulation results using the complete USC-SIPI Miscellaneous image dataset demonstrate that LSIC outperforms or reach state of the art suggested by many peer algorithms. All these analysis and results demonstrate that the LSIC is very suitable for digital image encryption. Finally, we open source the LSIC MATLAB code under webpage https://sites.google.com/site/tuftsyuewu/source-code., Comment: 26 pages, 17 figures, and 7 tables