Parallel finite automata is a formal model which can describe the complex control structure, but its composition model may exist redundant elements. This problem will increase the complexity of model. This paper resea...
详细信息
作者:
Huang, BaoyuMao, BoCollege of Information Engineering
Collaborative Innovation Center for Modern Grain Circulation and Safety Jiangsu Key Laboratory of Grain Big-data Mining Nanjing University of Finance Economics Nanjing China
Many applications demand simplified 3D building models for fast visualization and analysis purposes. However, due to the huge spatial data, it intensifies the load of the computer in the visualization process, and the...
详细信息
High-quality underwater images play an important role in obtaining and understanding underwater information. However, raw underwater images have many problems, such as low contrast, chromatic aberration, blur and low ...
High-quality underwater images play an important role in obtaining and understanding underwater information. However, raw underwater images have many problems, such as low contrast, chromatic aberration, blur and low light, which seriously restrict the progress of other underwater tasks. In this paper, we propose an Underwater Mixed Spatial Attention Network (UMSAN) for underwater image enhancement. The evaluation on the EUVP dataset and some other real underwater images demonstrates that our network is sufficient against several of the advanced models.
Current single image derain methods cannot solve the heavy rain situation well. In this paper, based on the physical model of a rainy image, we build a two-stage network, TSF-Net, which combines model-driven and data-...
Current single image derain methods cannot solve the heavy rain situation well. In this paper, based on the physical model of a rainy image, we build a two-stage network, TSF-Net, which combines model-driven and data-driven methods. The first stage gets the rain streaks, atmospheric light, and transmission map to obtain the coarse rain-free image by the physical model. The second stage is a fully convolutional neural network with the structure of U-Net. The proposed Multi-Scale Projection Fusion Block (MSPFB) module, which can perceive the spatial information across the scale of images, is employed to remove the residual rain and fuzzy parts in the first stage and obtain a refined clean image. Extensive experiments show that our TSF-Net achieves better accuracy and visual improvements against state-of-the-art methods.
Objective: This paper gives context on recent literature regarding the development of digital personal health libraries (PHL) and provides insights into the potential application of consumer health informatics in dive...
详细信息
作者:
Hong, S.Mao, B.College of Information Engineering
Collaborative Innovation Centre for Modern Grain Circulation and Safety Jiangsu Key Laboratory of Grain Big-data Mining Nanjing University of Finance and Economics Nanjing210023 China
The logistics industry plays a very important role in the operation of modern cities. Meanwhile, the development of logistics industry has derived various problems that are urgent to be solved, such as the safety of l...
详细信息
We study the adversarial multi-armed bandit problem, in which a player must iteratively make online decisions with linear loss vectors and hopes to achieve a small total loss. We consider a natural measure on the loss...
详细信息
Single image deraining task is still a very challenging task due to its ill-posed nature in reality. Recently, researchers have tried to fix this issue by training the CNN-based end-to-end models, but they still canno...
Single image deraining task is still a very challenging task due to its ill-posed nature in reality. Recently, researchers have tried to fix this issue by training the CNN-based end-to-end models, but they still cannot extract the negative rain streaks from rainy images precisely, which usually leads to an over de-rained or under de-rained result. To handle this issue, this paper proposes a new coarse-to-fine single image deraining framework termed Multi-stream Hybrid Deraining Network (shortly, MH-DerainNet). To obtain the negative rain streaks during training process more accurately, we present a new module named dual path residual dense block, i.e., Residual path and Dense path. The Residual path is used to reuse com-mon features from the previous layers while the Dense path can explore new features. In addition, to concatenate different scaled features, we also apply the idea of multi-stream with shortcuts between cascaded dual path residual dense block based streams. To obtain more distinct derained images, we combine the SSIM loss and perceptual loss to preserve the per-pixel similarity as well as preserving the global structures so that the deraining result is more accurate. Extensive experi-ments on both synthetic and real rainy images demonstrate that our MH-DerainNet can deliver significant improvements over several recent state-of-the-art methods.
The rapid development of information technology has involved advances in artificial intelligence(AI),big data processing,and cloud computing,with significant and farreaching effects on the structure and efficiency of ...
详细信息
The rapid development of information technology has involved advances in artificial intelligence(AI),big data processing,and cloud computing,with significant and farreaching effects on the structure and efficiency of the traditional healthcare industry,as well as the establishment and maintenance of modern medical management information *** solutions for handling data in the medical field,such as electronic medical records,medical imaging technology,medical big data,intelligent drug design,and smart health management systems have emerged,which improve the standardization and accuracy of clinical decision making,while providing more dimensions of data accumulation for medical knowledge-based *** developments can also support physicians and researchers in the optimization of treatment plans,and decision making about optimal treatment *** review aims to summarize recent advances in the research and clinical use of AI in pediatrics.
To overcome biomedical ontology’s heterogeneous problem, biomedical ontology matchers are developed to search for bridges of knowledge (or alignment) between heterogeneous biomedical ontologies. Since it is a complex...
详细信息
暂无评论