Search

IntroductionThe fusion of infrared and visible images can improve image quality and eliminate the impact of changes in the agricultural working environment on the information perception of intelligent agricultural systems.MethodsIn this paper, a distributed fusion architecture for infrared and visible image fusion is proposed, termed RADFNet, based on residual CNN (RDCNN), edge attention, and multiscale channel attention. The RDCNN-based network realizes image fusion through three channels. It employs a distributed fusion framework to make the most of the fusion output of the previous step. Two channels utilize residual modules with multiscale channel attention to extract the features from infrared and visible images, which are used for fusion in the other channel. Afterward, the extracted features and the fusion results from the previous step are fed to the fusion channel, which can reduce the loss in the target information from the infrared image and the texture information from the visible image. To improve the feature learning effect of the module and information quality in the fused image, we design two loss functions, namely, pixel strength with texture loss and structure similarity with texture loss.Results and discussionExtensive experimental results on public datasets demonstrate that our model has superior performance in improving the fusion quality and has achieved comparable results over the state-of-the-art image fusion algorithms in terms of visual effect and quantitative metrics.

As a new technology to reconfigure wireless communication environment by signal reflection controlled by software, intelligent reflecting surface (IRS) has attracted lots of attention in recent years. Compared with conventional relay system, the relay system aided by IRS can effectively reduce the cost and energy consumption, and significantly enhance the system performance. However, the phase quantization error generated by IRS with discrete phase shifter may degrade the receiving performance of the receiver. To analyze the performance loss caused by IRS phase quantization error, based on the law of large numbers and Rayleigh distribution, the closed-form expressions for the signal-to-noise ratio (SNR) performance loss and achievable rate of the IRS-aided amplify-and-forward (AF) relay network, which are related to the number of phase shifter quantization bits, are derived under the line-of-sight (LoS) channels and Rayleigh channels, respectively. Moreover, their approximate performance loss closed-form expressions are also derived based on the Taylor series expansion. Simulation results show that the performance losses of SNR and achievable rate decrease with the number of quantization bits increases gradually. When the number of quantization bits is larger than or equal to 3, the SNR performance loss of the system is smaller than 0.23dB, and the achievable rate loss is less than 0.04bits/s/Hz, regardless of the LoS channels or Rayleigh channels.

To protect the patient information in medical images, this article proposes a robust watermarking algorithm for medical images based on Harris-SURF-DCT. First, the corners of the medical image are extracted using the Harris corner detection algorithm, and then, the previously extracted corners are described using the method of describing feature points in the SURF algorithm to generate the feature descriptor matrix. Then, the feature descriptor matrix is processed through the perceptual hash algorithm to obtain the feature vector of the medical image, which is a binary feature vector with a size of 32 bits. Secondly, to enhance the security of the watermark information, the logistic map algorithm is used to encrypt the watermark before embedding the watermark. Finally, with the help of cryptography knowledge, third party, and zero-watermarking technology, the algorithm can embed the watermark without modifying the medical image. When extracting the watermark, the algorithm can extract the watermark from the test image without the original image. In addition, the algorithm has strong robustness to conventional attacks and geometric attacks. Especially under geometric attacks, the algorithm performs better.

This study aims to explore the factors that affect consumers’ behavior in adaptation and use of internet-based services (IBS) during the COVID-19 crisis. In this study, technology acceptance model (TAM) was applied to predict the behavioral intention of active social media users among the Pakistan population based on the revised model of the TAM model. And the data of external factors facilitating conditions (FC), social impact (SI), and task technology fit (TTF) were collected from active social media users of Pakistan by using structured questionnaires. After performing Pearson’s correlation and linear regression on the collected data, findings have shown that the outcome variable, i.e., behavioral intention, exhibited significant correlation with all variables except for perceived ease of use (PEoU). Further analysis revealed mixed results wherein FC and TTF can make a significant influence on perceived usefulness (PU) and PEoU, respectively. In addition, PU can significantly affect attitude (ATT) towards the use of IBS while the use of IBS has been affected by behavioral INT during the outbreak of COVID-19 in Pakistan.

Traditional service composition methods usually address the constraint-satisfied service composition (CSSC) problem with static web services. Such solutions however are inapplicable to the dynamic scenarios where the services or their QoS values may change over time. Some recent studies are proposed to use reinforcement learning, especially, integrate the idea of Q-learning, to solve the dynamic CSSC problem. However, such Q-learning algorithm relies on Q-table to search for optimal candidate services. When the problem of CSSC becomes complex, the number of states in Q-table is very large and the cost of the Q-learning model will become extremely high. In this paper, the authors propose a novel solution to address this issue. By training a DQN network to replace the Q-table, this solution can effectively model the uncertainty of services with fine-grained QoS attributes and choose suitable candidate services to compose on the fly in the dynamic scenarios. Experimental results on both artificial and real datasets demonstrate the effectiveness of the method.

Chest imaging plays an essential role in diagnosing and predicting patients with COVID-19 with evidence of worsening respiratory status. Many deep learning-based diagnostic models for pneumonia have been developed to enable computer-aided diagnosis. However, the long training and inference time make them inflexible. In addition, the lack of interpretability reduces their credibility in clinical medical practice. This paper presents CMT, a model with interpretability and rapid recognition of pneumonia, especially COVID-19 positive. Multiple convolutional layers in CMT are first used to extract features in CXR images, and then Transformer is applied to calculate the possibility of each symptom. To improve the model's generalization performance and to address the problem of sparse medical image data, we propose Feature Fusion Augmentation (FFA), a plug-and-play method for image augmentation. It fuses the features of the two images to varying degrees to produce a new image that does not deviate from the original distribution. Furthermore, to reduce the computational complexity and accelerate the convergence, we propose Multilevel Multi-Head Self-Attention (MMSA), which computes attention on different levels to establish the relationship between global and local features. It significantly improves the model performance while substantially reducing its training and inference time. Experimental results on the largest COVID-19 dataset show the proposed CMT has state-of-the-art performance. The effectiveness of FFA and MMSA is demonstrated in the ablation experiments. In addition, the weights and feature activation maps of the model inference process are visualized to show the CMT's interpretability., Comment: 12 pages, 7 figures

Plant play a very important role in the protection of the ecological balance. Compared to manual identification of plants, automated plant identification enable experts to process significantly greater numbers of plants with higher efficiencies in shorter periods of time. In this study, we propose an effective deep Convolutional Neural Network (CNN)-based model that is capable of automatically identifying and classifying plant species in Hainan by studying the details of their leaves. We apply transfer learning based on CNN to fine-tune the pre-trained models. Further, the optimal values of associated hyperparameters that maximize the accuracy of the proposed method are determined. Finally, experiments are carried out on two available botanical datasets: the Flavia dataset with 32 classes and the HNPlant dataset with 10 classes. The results demonstrate that the highest classification accuracies exhibited by the proposed CNN-based model on the Flavia and HNPlant datasets are 89% and 95%, respectively, thus establishing their effectiveness.

In this paper, the basic theory of geometric active contour model was studied, for the degradation problems of low contrast, fuzzy and color distortion of underwater image data. Combined with the problem of underwater image degradation caused by unevenness of light and water's scattering and absorption of light, the adaptive GACV image segmentation algorithm was improved, and a mathematical model of underwater image segmentation based on the improved adaptive GACV image segmentation algorithm was established, and the underwater images were emulated. The simulation results showed that the algorithm fully segmented underwater images with low contrast and fuzzy features, which achieved the expected segmentation effect and shows good robustness.

As an exchange foundation of cloud software services, the cloud marketplace plays an important role to promote the capability and creativity of cloud services. However, existing cloud marketplaces are usually based on online stores operated by some particular companies. Users are usually required to get involved in their particular ecosystems, seriously restricting their creativity and competition. Moreover, these marketplaces lack efficient and flexible service publishing and discovery mechanisms. Therefore, this makes them difficult to handle a huge number of dynamic cloud service exchanges efficiently. To address this issue, we propose in this paper an open cloud marketplace middleware system for cloud software services based on a service publishing and subscription model. This model extends traditional content-based distributed publish/subscribe paradigm. By introducing the matching mechanism of multidimensional contents such as service functionality, service behavior, and quality of service etc, a new distributed publish/subscribe paradigm for service publishing and subscription is proposed. This new paradigm allows service providers to publish their services to and allows service consumers to subscribe services from the open marketplace based on service descriptions and requirements. We illustrate the design of the open cloud marketplace and present our preliminary results.