Search

Cancer region detection (CRD) and subtyping are two fundamental tasks in digital pathology image analysis. The development of data-driven models for CRD and subtyping on whole-slide images (WSIs) would mitigate the burden of pathologists and improve their accuracy in diagnosis. However, the existing models are facing two major limitations. Firstly, they typically require large-scale datasets with precise annotations, which contradicts with the original intention of reducing labor effort. Secondly, for the subtyping task, the non-cancerous regions are treated as the same as cancerous regions within a WSI, which confuses a subtyping model in its training process. To tackle the latter limitation, the previous research proposed to perform CRD first for ruling out the non-cancerous region, then train a subtyping model based on the remaining cancerous patches. However, separately training ignores the interaction of these two tasks, also leads to propagating the error of the CRD task to the subtyping task. To address these issues and concurrently improve the performance on both CRD and subtyping tasks, we propose a semi-supervised multi-task learning (MTL) framework for cancer classification. Our framework consists of a backbone feature extractor, two task-specific classifiers, and a weight control mechanism. The backbone feature extractor is shared by two task-specific classifiers, such that the interaction of CRD and subtyping tasks can be captured. The weight control mechanism preserves the sequential relationship of these two tasks and guarantees the error back-propagation from the subtyping task to the CRD task under the MTL framework. We train the overall framework in a semi-supervised setting, where datasets only involve small quantities of annotations produced by our minimal point-based (min-point) annotation strategy. Extensive experiments on four large datasets with different cancer types demonstrate the effectiveness of the proposed framework in both accuracy and generalization.

With the rapid development of Internet technology, more and more people begin to pay attention to and study the information mining and data analysis in social networks, and the maximization of influence widely used in marketing have gradually become the focus of people's research. Nowadays, most of the researches on the maximization of influence under topic preference in social network information dissemination only tend to judge the influence of nodes on a certain information from the interaction of nodes to the topic content of information. The combination of the topology of social networks and the topic content of information dissemination is not considered. In view of this situation, an influence maximization algorithm under topic preference(IMATP algorithm) is proposed, which evaluates the comprehensive influence of nodes by combining node centrality and theme influence. Experimental results on data sets of different sizes from Twitter show that the algorithm proposed in this paper is closer to the actual propagation situation. And compared with the classical algorithm, this algorithm not only improves the scope of influence, but also improves the running efficiency.

In recent years, Reversible Data Hidden (RDH) has been paid more and more attention by researchers. Existing RDH algorithms based on histogram shifting require that there should be no overlap between the multiple pairs of maximum and minimum points, so that the embedding capacity is relatively small. In order to improve the embedding capacity, we propose a histogram shifting technique which allows the region between the multiple pairs of maximum and minimum points to be overlapped. This scheme simplifies the process of selecting frequency points. It only needs to find some maximum points and minimum points, and one maximum point matches another minimum point according to their distance. For each pair of maximum and minimum point, histogram shifting is applied to embed the secret information. The experimental results show that the embedding capacity is greatly improved while the image quality reduces slightly.

Here we develop a simple and effective nose/tongue sensor array based on Co-N-C single-atom nanozymes-3,3′,5,5′-tetramethylbenzidine (TMB)-H2O2 for colorimetric discrimination of antioxidants, which makes use of the color reaction of TMB oxidation by H2O2 in two different pH (3.8 and 4.6) environments under the catalysis of Co-N-C nanoenzyme with peroxidase-like activity. Different antioxidants have varying reducing ability to the oxidation products of TMB (oxTMB), thus resulting in differential absorbance and color changes. Linear discriminant analysis (LDA) results indicate that the sensor array successfully identified 7 antioxidants, i.e., glutathione (GSH), ascorbic acid (AA), cysteine (Cys), tannin (TA), Catechin (C), dopamine (DA), and uric acid (UA) in both buffer and even serum samples. Additionally, the performance of the sensor array was validated with antioxidant mixtures, individual antioxidants with different concentrations, and target antioxidants and interfering substances. In general, the versatile sensor array based on Co-N-C single-atom nanozymes provides an excellent strategy for identifying a variety of antioxidants, which exhibits a broad application prospect in medical diagnosis.

The Shack–Hartmann wavefront sensor (SHWFS) has been used to measure wavefronts. However, the dynamic range and measurement accuracy limits the application of the SHWFS. A new centroid estimation algorithm based on image segmentation and related techniques to expand the dynamic range based on search strategy is presented in this paper. In this study, each spot is segmented from a spot array image and the centroid of each spot is calculated only in regions of segmented spots. To compute the slope between the spot centroid and the sub-aperture centre, the calculated centroids are assigned to their corresponding sub-apertures using a search method. The proposed method overcomes the limitations that the centroid must be calculated in its corresponding detection area and reduces the measurement errors caused by noise, expanding the dynamic range significantly without any change to the SHWFS setup as well as improving the measurement accuracy under strong noise. Moreover, the proposed method provides a unit linear coefficient response of centroid estimation. The dynamic range for the proposed method is evaluated by numerical simulations and experiments. Applications include measurement accuracy and linearity.

Mg Al CO3 layered double hydroxide (LDH) was prepared via the hydrothermal method, and used to remove Congo red (CR) from the aqueous solution with the assist of ultrasound. The influence of ultrasonic power on the CR adsorption, together with adsorption kinetics and isotherms were investigated. The mechanism of ultrasound-assisted adsorption was elucidated. The results showed that compared with stirring and shaking, ultrasound-assisted adsorption resulted in a significantly shorter adsorption equilibrium time of 120 min under initial 0.2 g/L CR and 0.2 g/L LDH dosage. The optimal ultrasonic power was 180 W at which a maximum adsorption capacity of 934.43 mg/g was achieved. The adsorption kinetics fitted the pseudo-second order kinetic model well (R2 = 0.9999), and the adsorption isotherms fitted the Langmuir isotherm model well (R2 = 0.998). The adsorption rate remained above 60% after three cycles.

In order to enhance the performance of supercapacitors, electrode materials and electrolytes have drawn a great deal of attention, however, there is limited research on conductive agents. Carbon onions were reported as a promising conductive agent for supercapacitor. In this paper, highly graphitized carbon onions were prepared by chemical vapor deposition using methane as carbon source and then purified with pickling - calcination - pickling method. The purified carbon onions and commercial super-P were employed as conductive agents of supercapacitors with the same proportion and dosage. Electrochemical tests reveal that carbon onion conductive agent provides an excellent electrochemical performance for supercapacitor, including high specific capacity (142.3 F/g) and relatively stable cycle stability (86.9% capacity retention after 5000 cycles). We believe that the nanopores among packing carbon onions are responsible for the enhanced performance of supercapacitor.

The novel magnetically retrievable Bi2WO6/MCNOs (MCNOs, magnetic carbon nano-onions) composite was synthesized via a simple hydrothermal method. The characteristics of the as-prepared materials were explored by XRD, Raman, SEM, TEM, XPS, UV-vis DRS, N2 adsorption-desorption techniques and FT-IR techniques, and the paramagnetic nature of the as-prepared composite was determined by hysteresis loops measurements. In this work, rhodamine B (RhB), methyl orange (MO), methylene blue (MB), tetracycline (TC) and p-nitrophenol (PNP) were chosen as the typical organic pollutants to evaluate the photodegradation efficiency of the as-prepared composite. The results indicated that Bi2WO6/MCNOs have a superior performance on photocatalytic activity than pure Bi2WO6. Furthermore, a possible mechanism was proposed based on the physic-chemical and photocatalytic properties. This work will provide new methodology for designing next generation photocatalyst for environmental purification.

Alzheimer’s disease (AD) is one of the most common neurodegenerative diseases in the elderly. The aim of this study was to explore the effects of AD on cardiac function and autonomic nervous function, and the feasibility of electrocardiogram (ECG) in monitoring the development of AD. APP/PS1 double transgenic mice were used in the Morris water maze (MWM) experiment to evaluate the changes of cognitive ability of AD mice, then the non-invasive ECG acquisition system was used and the changes of ECG intervals and heart rate variability (HRV) were analyzed. AD mice already had cognitive dysfunction at the age of 5 months, reaching the level of mild dementia, and the degree of dementia increased with the course of disease. There were no significant changes in ECG intervals in the AD group at each month. The mean square of successive RR interval differences, percentage of intervals >6 ms different from preceding interval, and normalized high frequency power component in the AD group were decreased and low-to-high frequency power ratio and normalized low frequency power component were increased. Combined with the results of the MWM, it was shown that the regulation mechanism of sympathetic and parasympathetic nerves in mice was already imbalanced in early stage AD, which was manifested as the increase of excessive activity of sympathetic nerves and the inhibition of parasympathetic activities. Therefore, ECG-based analysis of HRV may become a means of daily monitoring of AD and provide an auxiliary basis for clinical diagnosis.

Histological subtype of papillary (p) renal cellular and cell-layer level patterns almost cannot be captured by existing CNN-based models in large-size histopathological images, which brings obstacles to directly applying these models to such a fine-grained classification task. This paper proposes a novel instance-based Vision Transformer (i-ViT) to learn robust representations of histopathological images for the pRCC subtyping task by extracting finer features from instance patches (by cropping around segmented nuclei and assigning predicted grades). The proposed i-ViT takes top-K instances as the inputs and aggregates them for capturing both the cellular and cell-layer level patterns by a position-embedding layer, a grade-embedding layer, and a multi-head multi-layer self-attention module. To evaluate the performance of the proposed framework, experienced pathologists select 1162 regions of interest from 171 whole slide images of type 1 and type 2 pRCC. Experimental results show that the proposed method achieves better performance than existing CNN-based models with a significant margin.

In this paper, we report on a non-directed line-of-sight (ND-LOS) visible light communication (VLC) system using an automated rotating photodetector array receiver with a wide field-of-view, which is more flexible in practical applications. The geometry of ND-LOS propagation model is numerically analyzed using Monte Carlo method. We experimentally demonstrate a 180 Mbps real-time on-off keying ND-LOS VLC link with a bit error rate of less than 2.87 ×10−6 over a transmission distance of 2.3 m, which employs a single white phosphorous light-emitting diode with a transmit power of 1 W.

Ultraviolet is a vital spectrum in wireless optical communication. There are a lot of studies have been done in wireless ultraviolet communication. However, Ultraviolet communication networking mainly focuses on self-organizing networks and multi-hop networks. We proposed a new multi-access scheme of star topology network for confidential communications in semi-open spaces on the surface of ships. We verified the feasibility of the scheme through theoretical analysis and system simulation. It is proved that the scheme is suitable for ultraviolet communication and has better performance in a multi-user network. This scheme could be a sub-scheme for the next generation communication network combined with ultraviolet light communication.

BACKGROUND Current time series models such as Long Short-Term Memory Network (LSTM), Autoregressive Integrated Moving Average Model (ARIMA), and machine learning ensembles have been used to forecast the number of novel daily COVID-19 cases by specific region. The most common are ARIMA models. Lacking in the models are explainability by intuitively showing why and how the forecasts are being made, and the ability to generalize the models to be used in any country or region. Other models that use more novel approaches, lack in ease of use and access to data. OBJECTIVE To fill this gap by providing a robust method that is explainable, generic, and easy to use. METHODS The presented model, Elastic Net COVID-19 Forecaster or EN-CoF for short, uses a linear regressor trained using elastic net to learn meaningful weights of 17 previous 5-day averages to forecast the next 5-day average for any country. RESULTS EN-CoF made novel predictions for 124 countries, up to two weeks in advance, with a coefficient of determination, r2, value of 0.9860. For comparisons, the most commonly use model ARIMA was used on the same data and scored a comparable r2 value of 0.9872. CONCLUSIONS EN-CoF results, r2 values, are on par with state-of-the-art methods like ARIMA. The advantages of EN-CoF is that it is generic enough to be used for any country or region, provides intuitive static weights making it more explainable and easier to apply, and can be used without extensive machine learning or other technical knowledge.

Potholes are a structural damage to the road with hollow which can cause severe traffic accidents and impact road efficiency. In this paper, we propose an efficient pothole detection system using deep learning algorithms which can detect potholes on the road automatically. Four models are trained and tested with preprocessed dataset, including YOLO V3, SSD, HOG with SVM and Faster R-CNN. In the phase one, initial images with potholes and non-potholes are collected and labeled. In the phase two, the four models are trained and tested for the accuracy and loss comparison with the processed image dataset. Finally, the accuracy and performance of all four models are analyzed. The experimental results show that the YOLO V3 model performs best for its faster and more reliable detection results.

As an effective image preprocessing method, color correction is widely used in the field of image stitching. However, due to the changeable visual environment, parallax often occurs in mosaic images. So far, there are few algorithms dealing with color correction under parallax situation. And most color correction methods produce poor results in images with parallax. This paper introduces a color correction algorithm which can also be applied in the case of parallax in image mosaic. We use VFC algorithm, color hue information and iterative calculation to obtain better color correction results. Compared with existing algorithms, it has the wider range of applications. The experiments show that our method has faster time efficiency and better results.

Pathological is crucial to cancer diagnosis. Usually, Pathologists draw their conclusion based on observed cell and tissue structure on histology slides. Rapid development in machine learning, especially deep learning have established robust and accurate classifiers. They are being used to analyze histopathological slides and assist pathologists in diagnosis. Most machine learning systems rely heavily on annotated data sets to gain experiences and knowledge to correctly and accurately perform various tasks such as classification and segmentation. This work investigates different granularity of annotations in histopathological data set including image-wise, bounding box, ellipse-wise, and pixel-wise to verify the influence of annotation in pathological slide on deep learning models. We design corresponding experiments to test classification and segmentation performance of deep learning models based on annotations with different annotation granularity. In classification, state-of-the-art deep learning-based classifiers perform better when trained by pixel-wise annotation dataset. On average, precision, recall and F1-score improves by 7.87%, 8.83% and 7.85% respectively. Thus, it is suggested that finer granularity annotations are better utilized by deep learning algorithms in classification tasks. Similarly, semantic segmentation algorithms can achieve 8.33% better segmentation accuracy when trained by pixel-wise annotations. Our study shows not only that finer-grained annotation can improve the performance of deep learning models, but also help extracts more accurate phenotypic information from histopathological slides. Intelligence systems trained on granular annotations may help pathologists inspecting certain regions for better diagnosis. The compartmentalized prediction approach similar to this work may contribute to phenotype and genotype association studies., Comment: Accepted by AIPath2019 Workshop in BIBM2019. 7 pages, 4 figures, 4 tables

Cancerous region detection and subtyping in whole-slide images (WSIs) are fundamental for renal cell carcinoma (RCC) diagnosis. The main challenge in the development of automated RCC diagnostic systems is the lack of large-scale datasets with precise annotations. In this paper, we propose a framework that employs a semi-supervised learning (SSL) method to accurately detect cancerous regions with a novel annotation method called Minimal Point-Based (Min-Point) annotation. The predicted results are efficiently utilized by a hybrid loss training strategy in a classification model for subtyping. The annotator only needs to mark a few cancerous and non-cancerous points in each WSI. The experiments on three significant subtypes of RCC proved that the performance of the cancerous region detector trained with the Min-Point annotated dataset is comparable to the classifiers trained on the dataset with full cancerous region delineation. In subtyping, the proposed model outperforms the model trained with only whole-slide diagnostic labels by 12% in terms of the testing f1-score. We believe that our “detect then classify” schema combined with the Min-Point annotation would set a standard for developing intelligent systems with similar challenges.

The global market size of conversational assistants (chatbots) is expected to grow to USD 9.4 billion by 2024, according to MarketsandMarkets. Despite the wide use of chatbots, leakage of personal information through chatbots poses serious privacy concerns for their users. In this work, we propose to protect personal information by warning users of detected suspicious sentences generated by conversational assistants. The detection task is formulated as an alignment optimization problem and a new dataset PERSONA-LEAKAGE is collected for evaluation. In this paper, we propose two novel constrained alignment models, which consistently outperform baseline methods on Moreover, we conduct analysis on the behavior of recently proposed personalized chit-chat dialogue systems. The empirical results show that those systems suffer more from personal information disclosure than the widely used Seq2Seq model and the language model. In those cases, a significant number of information leaking utterances can be detected by our models with high precision.

Transition from conventional to digital pathology requires a new category of biomedical informatic infrastructure which could facilitate delicate pathological routine. Pathological diagnoses are sensitive to many external factors and is known to be subjective. Only systems that can meet strict requirements in pathology would be able to run along pathological routines and eventually digitized the study area, and the developed platform should comply with existing pathological routines and international standards. Currently, there are a number of available software tools which can perform histopathological tasks including virtual slide viewing, annotating, and basic image analysis, however, none of them can serve as a digital platform for pathology. Here we describe OpenHI2, an enhanced version Open Histopathological Image platform which is capable of supporting all basic pathological tasks and file formats; ready to be deployed in medical institutions on a standard server environment or cloud computing infrastructure. In this paper, we also describe the development decisions for the platform and propose solutions to overcome technical challenges so that OpenHI2 could be used as a platform for histopathological images. Further addition can be made to the platform since each component is modularized and fully documented. OpenHI2 is free, open-source, and available at https://gitlab.com/BioAI/OpenHI., Comment: Preprint version accepted to AIPath2019 workshop at BIBM2019. 6 pages, 3 figures, 2 tables

Background Sleep affects almost all aspects, including health, memory and quality of life. It is important to distinguish wake/sleep correctly in sleep monitoring. Supervised recognition algorithms are trained on expensive and laborious polysomnography (PSG), and individual differences can’t be ignored. Most of the unsupervised recognition algorithms are based on acceleration signal, and body movement seriously affects its accuracy. Objective In order to improve the generalization ability of wake/sleep recognition and avoid the adverse effects of body movement, we propose an unsupervised method that only needs heart rate variability (HRV) to recognize wake/sleep states, and verify the classification results in a public database. Methods Shapelets algorithm is used to quantify the similarity between HRV segments, and K-means clustering algorithm is used to improve the shapelets algorithm to realize the classification of wake/sleep. The sleep tags in the database are used to verify and compare the classification results before and after the improvement. Besides, the influence of unbalanced samples on the classification performance of the two algorithms is analyzed. Results The accuracy of the combined shapelets and K-means (SLKM) algorithm is 0.7800 ± 0.0692, 0.8826 ± 0.0533 in two databases. Compared with the shapelets algorithm, the accuracy is improved by 11.47% and 15.49% respectively. Conclusions This method can effectively realize wake/sleep recognition based on individual HRV. It has stronger robustness, which is very suitable for large-scale and long-time sleep monitoring.

This paper proposes a new type of composite shear wall with concrete-filled steel tubular frames and artificial vertical slits. The seismic behaviour of this new shear wall is studied by using a cyclic loading test. The failure mode, load-carrying capacity, ductility, stiffness degradation, hysteresis behaviour and energy dissipating capacity shown in the course of the test are studied. And this paper proposes the calculation formula of normal section bearing capacity. The main research results are as follows. Firstly, the failure mode of this new shear wall is the flexural failure. Secondly, compared with the ordinary reinforced concrete shear wall with slits, the new shear walls with concrete-filled steel tubular frames and slits show a better performance in load-carrying capacity, ductility, stiffness degradation and energy dissipating capacity.

Hyperspectral images (HSIs) provide invaluable information in both spectral and spatial domains for image classification tasks. In this paper, we use semantic representation as a middle-level feature to describe image pixels’ characteristics. Deriving effective semantic representation is critical for achieving good classification performance. Since different image descriptors depict characteristics from different perspectives, combining multiple features in the same semantic space makes semantic representation more meaningful. First, a probabilistic support vector machine is used to generate semantic representation-based multifeatures. In order to derive better semantic representation, we introduce a new adaptive spatial regularizer that well exploits the local spatial information, while a nonlocal regularizer is also used to search for global patch-pair similarities in the whole image. We combine multiple features with local and nonlocal spatial constraints using an extended Markov random field model in the semantic space. Experimental results on three hyperspectral data sets show that the proposed method provides better performance than several state-of-the-art techniques in terms of region uniformity, overall accuracy, average accuracy, and Kappa statistics.

Lodging is a major yield-limiting factor in wheat, so we need a method to collect wheat lodging information efficiently. Traditional extraction of crop lodging is cost intensive and time consuming; satellite remote sensing technology is vulnerable to the cloud and is limited by its temporal and spatial resolution. Therefore, a new method for wheat lodging extraction based on a watershed algorithm combined with an adaptive threshold segmentation hybrid algorithm is proposed. An unmanned aerial vehicle equipped with a high-resolution thermal infrared and visible sensor was used to acquire images of lodged and nonlodged wheat. Results indicate that there are morphological differences between lodged and nonlodged wheat and that the spectral reflectance and temperature of lodged wheat is higher than that of nonlodged wheat. The watershed algorithm is sensitive to the edge response and can reflect the contour characteristics of the lodging area. However, the impact of noise is obvious; there are salt-and-pepper patches. The wheat lodging ratios were calculated as 17.81%, with an accuracy of 80.26%. With adaptive threshold segmentation, the wheat lodging ratios were calculated as 18.78%, with an accuracy of 84.65%. If these two methods are fused, the influence of noise can be reduced, and an extraction accuracy of 93.58% can be achieved. This research provides a new idea for highly efficient crop lodging data collection, which can be used for precision farming and the selection of crop genotypes with high lodging resistance.

Aberrations degrade the accuracy of quantitative, imaging-based measurements, like particle image velocimetry (PIV). Adaptive optical elements can in principle correct the wavefront distortions, but are limited by their technical specifications. Here we propose an actuator-free correction based on a multiple-input deep convolutional neural network which uses an additional input from a wavefront sensor to correct time-varying distortions. It is applied for imaging flow velocimetry to conduct measurements through a fluctuating air-water phase boundary. Dataset for neural network is generated by an experimental setup with a deformable mirror. Correction performance of trained model is estimated in terms of image quality, which is improved significantly, and flow measurement results, where the errors induced by the distortion from fluctuating phase boundary can be corrected by 82 %. The technique has the potential to replace classical closed-loop adaptive optical systems where the performance of the actuators is not sufficient.

In the field of image processing, color correction is usually used when there are significant differences in tone or brightness in the original images. This method is also often used as an image preprocessing method to assist many other applications. This paper introduces a method of color correction using rational function model and histogram. And we use two different mapping methods to get best stitching results according to different application scenarios. The problem of color discontinuity in many color correction algorithms is solved, and the experimental results show that our method has a faster time efficiency.

Human dynamic behavior detection and classification based on video plays an important role in understanding human motion, and it has many interesting applications in crime watch, security surveillance monitoring, sports, and so on. This paper reports our initial research work on dynamic human behavior pattern detection and classification based on real-time video data. We focus on four types of detailed dynamic human behaviors: walking, standing, running, and sitting. A Convolutional Long Short-Term Networks (CLSTM) model is proposed and used for dynamic human behavior pattern detection and classification based on videos. This model combines the CNN model and LSTM to support learning, detection, and classification of dynamic human behavior patterns. AlexNet, one of CNN traditional architecture, is firstly used to learn visual representation of human behaviors based on time-based images. Later, the results will be used as inputs for the LSTM model to learn time-based sequence features to support detailed behavior pattern classification. A prototype of a Human Behavior Detection System is reported based on the CLSTM model, and some earlier case study results are presented.

Artificial intelligence applications provide tremendous opportunities to improve human life and drive innovation. AI systems/applications which operate in a real-world environment have to encounter an infinite set of feasible scenarios. Conventional testing approach to test the AI application allows only limited testing and does not allow taking the different contexts into consideration and may lead to insufficient validation and characterization. Therefore, to ensure robustness, certainty and reliability of AI applications, the authors applied classification-based AI software testing framework and 3D decision tables to generate test cases. Moreover, the authors compared the quality assurance metrics (accuracy, correctness, reliability and consistency) of AI and non-AI functions in the AI mobile application scenario. Our results indicate and confirm that complete AI function validation is not possible with conventional testing methods, but AI software testing strategy proposed based on classification framework and 3D decision tables has a good effect.

This paper presents a new spectral–spatial feature learning method for hyperspectral image classification, which integrates spectral and spatial information into group sparse coding (GSC) via clusters, each of which is an adaptive spatial partition of pixels. The clusters derived from the segmentation maps by the mean-shift algorithm are regarded as groups in GSC, where pixels within the same group are simultaneously represented by a sparse linear combination of a few common atoms in a given dictionary, thus enforcing spatial smoothness across the pixels in the same segmentation region to learn a spectral–spatial joint sparse representation. Finally, the recovered sparse representation can be viewed as a new feature and used directly for classification (e.g., by support vector machine). In comparison with other spectral–spatial classification techniques that exploit a fixed neighborhood system and force neighboring pixels to share a common sparsity pattern, the proposed method is more flexible and able to obtain adaptive spatial neighborhood correlations for spectral–spatial joint sparse coding. In addition, we also develop kernel GSC (KGSC) by incorporating the kernel trick into GSC to capture nonlinear relationships. The developed KGSC can also be applied to learning kernel sparse representation under the framework of the proposed spectral–spatial method, leading to a new spectral–spatial kernel sparse representation algorithm. Experimental results on three real hyperspectral datasets indicate that the proposed methods improve classification accuracy and provide distinctive classification maps, especially at small regions and boundaries in an image, compared with other similar approaches.

The piles of channel bridges generally exist in the form of steel casing composite piles (SCC pile, namely, the reinforced concrete pile wrapped by steel casing) owing to construction factors, such as drilling and temporary platforms. To investigate the effect of steel casings on the behaviors of the original reinforced concrete (RC) piles, and to clarify the reasons for this effect, we conducted a lateral loading experiment on three sets of piles. The three sets of piles included a set of RC piles, a set of SCC piles with partially set steel casings, and a set of piles with fully set steel casings. Each set comprised two identical specimens. In addition, to provide a theoretical basis for determining appropriate parameters for steel casings in practice, parameter analysis was conducted by numerical modeling. The results indicated that the contact force between the steel casing and the RC shaft directly caused the transformation of failure mode of the pile. When the length of steel casing was insufficient, the accumulated contact force resulted in great shear force and bending moment in the uncased RC shaft, and the failure mode of the pile would transform from bending to bending-shear failure. The mechanical behaviors and failure modes of SCC piles were significantly affected by the relative position between the steel casing and the equivalent plastic zone of the original RC shaft. Only if the equivalent plastic zone of the original RC shaft was wrapped by the steel casing, the critical state would be governed by the steel casing, and the bearing capacity and toughness of SCC pile would be significantly improved. Moreover, a fitting formula of the optimum length-to-diameter ratio for a steel casing was proposed, and the most unfavorable length-to-diameter ratio was discussed.

The novel magnetically retrievable Bi

Web image analysis has witnessed an AI renaissance. The ILSVRC benchmark has been instrumental in providing a corpus and standardized evaluation. The NVIDIA AI City Challenge is envisioned to provide similar impetus to the analysis of image and video data that helps make cities smarter and safer. In its first year, this Challenge has focused on traffic video data. While millions of traffic video cameras around the world capture data, albeit low-quality, very little automated analysis and value creation results. Lack of labeled data, and trained models that can be deployed at the edge of the city fabric, ensure that most traffic video data goes through little or no automated analysis. Real-time and batch analysis of this data can provide vital breakthroughs in real-time traffic management as well as pedestrian safety. The NVIDIA AI City Challenge brought together 29 teams from universities in 4 continents to collaboratively annotate a 125 hour data set and then compete on detection, localization and classification tasks as well as traffic and safety application analytics tasks. The result is the largest high quality annotated data set, a set of models trained using NVIDIA AI City Edge to Cloud platform and ready to be deployed at the edge solving traffic and safety problems for cities worldwide.

Water quality becomes one of the important quality factors for the quality life in smart cities. Recently, water quality has been degraded due to diverse forms of pollution caused by disposal of human wastes, industrial wastes, automobile wastes. The increasing pollution affects water quality and the quality of people's life. Hence, water quality evaluation, monitoring, and prediction become an important and hot research subject. In the past, many environmental researchers have dedicated their research efforts on this subject using conventional approaches. Recently, many researchers begin to use the big data analytics approach to studying, evaluating, and predicting water quality due to the advances of big data applications and the availability of environmental sensing networks and sensor data. This paper reviews the published research results relating to water quality evaluation and prediction. Moreover, the paper classifies and compares the applied big data analytics approaches and big data based prediction models for water quality assessment. Furthermore, the paper also discusses the future research needs and challenges.

This paper presents a novel method for hyperspectral classification combining multiple features and exploiting spatial information at the same time. We proposed a supervised classification method under the Markov random field (MRF)-based framework. Firstly using the probability SVM to map multiple features from different low-level subspace to the same semantic space (probability space), then integrating these features in semantic space with MRF-based model to enforce a smooth and accurate representation, in addition the manifold distance has been used in MRF-based model to measure the similarity of two point. To further improve the classification accuracy, a new approach of building the adaptive neighborhood has been proposed and used in our method. As our model is a derivable and convex problem, gradient descent can be used to solve this problem with less computational and time cost. Experimental results on real hyperspectral dataset shows that the proposed method provides improved classification accuracy in terms of the overall accuracy, average accuracy and kappa statistic.

The fast deployment of wireless networks and mobile technologies and the significant increase in the number of mobile device users have created a very strong demand for new innovative mobile commerce applications and services. Since mobile advertising is one of the important subjects in mobile commerce, studying and developing mobile advertising solutions becomes a hot research and business topic. This paper first discusses the basic concepts of mobile advertising. Then, it reports a mobile advertising system (known as SmartMobile-AD), including its supporting business process, functions, system architecture, and built-in technical solutions. The system supports the wireless advertising workflow for wireless service companies and publishers (or portals) to manage and deliver multimedia mobile ads to mobile users over the wireless Internet. The paper provides insights and design solutions on mobile advertisement targeting, delivery, tracking, and performance measurement. Furthermore, the paper presents application examples of the system.

Cloud computing introduces a new paradigm for software deployment, hosting, and service renting. Based on the XaaS architecture, a large number of users may share computing resources, platform services, and application software in a multi-tenancy approach. To ensure service availability, the system needs to support an advanced level of massive scalability so that it can provide necessary resources on demand following the pay-per-use pricing model. This chapter analyzes the unique requirements of cloud performance and scalability, compared with traditional distributed systems. Measurements are proposed with performance indicators, meters, and metrics identified from different perspectives. To support scalability testing in a dynamic environment, an agent-based testing framework is proposed to facilitate adaptive load generation and simulation using a two-layer control architecture.

Rapidly developing wireless net-working technology and the growing mobile-device user base have fueled interest in activities that deliver advertisements to mobile devices over a wireless network. Studies by wireless media research companies indicate that delivering permission-based alerts to wireless phones captures consumers' attention, drives response actions, and builds brand awareness. Wireless devices are accessible, personal, and location aware. These characteristics allow for highly targeted, flexible, and dynamic wireless advertisements. Yet the target audience is vast, and users must be able to search for information, issue inquiries, and make purchases at any mobile location.

User authentication is always the first step of security services/mechanisms in computing infrastructures. However, mobile devices and clouds bring challenges and also new dimensions into this topic, such as limited computing power, migrating objects, and various embedded sensors. When a resource located either in a mobile device or in a cloud is requested, associated security policies will trigger authentication procedures for compliance enforcement. This paper proposes a new multimodal approach, called NemoAuth, to mobile device user authentication. It has unique mnemonic and other interesting benefits, and also improves both security and user experience which breaks the limit of recent user authentication technologies.

The fast deployment of wireless networks and mobile technologies and the significant increase in the number of mobile device users have created a very strong demand for new innovative mobile commerce applications and services. Since mobile advertising is one of the important subjects in mobile commerce, studying and developing mobile advertising solutions becomes a hot research and business topic. This paper first discusses the basic concepts of mobile advertising. Then, it reports a mobile advertising system (known as SmartMobile-AD), including its supporting business process, functions, system architecture, and built-in technical solutions. The system supports the wireless advertising workflow for wireless service companies and publishers (or portals) to manage and deliver multimedia mobile ads to mobile users over the wireless Internet. The paper provides insights and design solutions on mobile advertisement targeting, delivery, tracking, and performance measurement. Furthermore, the paper presents application examples of the system.

The wide deployment of wireless networks and mobile technologies and the significant increase in the number of mobile device users have created a very strong demand for emerging mobile commerce applications and services. Wireless advertising is one of the important and hot subjects in m-commerce and wireless application. This paper reports a wireless advertising system (SmartMobile-AD) in terms of its supporting business process, functions, system architecture, and built-in technical solutions. The system supports the wireless advertising workflow for wireless service companies and publishers (or portals) to manage and deliver multimedia mobile ads to mobile users over the wireless Internet. The paper provides the insights on mobile ad targeting, delivery, tracking, and performance measurement. Furthermore, the paper presents the application examples of the system.