Search

AI technology has become a new creative tool, which brings infinite possibilities for painting creation. In this paper, we first parameterize the design of the painting through SolidWorks software and construct a semantic text with the design parameters obtained. Secondly, the text is inputted into the text-image similarity model based on a graph convolutional network, and the Stanford CoreNLP tool is used to perform dependency parsing on the input text and feature extraction on the textual semantic and spatial relation maps through two GCN networks. Finally, the visual parameters and visual comfort of the AI drawing samples are examined. The results show that the maximum difference between the cumulative scores comparing the reference image and the AI output image is 0.41, and this value is less than 1, which indicates that the AI drawing achieves the visual effect of quantifying the image quality. The efficiency and quality of drawing output can be improved through this study, which has both theoretical and practical significance.

This study develops an unstructured mesh generation tool for land surface modeling using a multi‐scale hexagon discrete mesh. The tool can automatically determine the required mesh resolution for different regions based on multi‐objective criteria such as elevation, slope, land cover, and land use. The refined unstructured meshes demonstrate significant enhancement in the representation of spatial heterogeneity. The tool is coupled with the Common Land Model (CoLM) to enable land surface simulations using unstructured meshes. Evaluations focused on runoff, river discharge, and inundation indicate improved model performance compared to traditional structured mesh‐based CoLM simulations under the same computational cost constraints. This tool provides new capabilities for more efficiently capturing localized land surface processes and extreme events. Plain Language Summary: The traditional square mesh cells used in land surface models (LSMs) have limitations in capturing the hydrological processes in regions with spatial solid heterogeneity, for example, mountains and forest‐grass transition zones. This study develops a new tool to generate unstructured mesh with flexible cell shapes and sizes. The tool can automatically adjust the mesh resolution based on the various land surface features. Thus, it can represent better spatial differences in elevation, vegetation, soil, etc. For example, it can use finer meshes for mountains and coarser meshes for desert areas. The new unstructured meshes are coupled with an existing LSM called Common Land Model. Model testing shows improved runoff, river discharge, and flooding simulation compared to traditional square mesh models under the same computation cost. Overall, the new unstructured mesh modeling approach enables more efficient and accurate simulation of localized land processes and extreme events, for example, droughts and floods. It has significance for disaster prevention and environmental studies. Key Points: An unstructured mesh generation tool for land surface models can automatically refine resolution based on multi‐objective landscape criteriaCoupling refined unstructured meshes with the Common Land Model (CoLM) model improves local hydrological simulation compared to traditional mesh‐based CoLMThe tool enables more efficient and accurate modeling of localized land surface processes and extreme events such as droughts and floods [ABSTRACT FROM AUTHOR]

Summary: As the development of deep learning and the continuous improvement of computing power, as well as the needs of social production, target detection has become a research hotspot in recent years. However, target detection algorithm has the problem that it is more sensitive to large targets and does not consider the feature‐feature interrelationship, which leads to a high false detection or missed detection rate of small targets. An small target detection method (C‐SSD) based on improved SSD is proposed, that replaces the backbone network VGG‐16 of the SSD network with the improved dense convolution network (C‐DenseNet) network to achieves further feature fusion through fast connections between dense blocks. The Introduction of residuals in the prediction layer and DIoU‐NMS further improves the detection accuracy. Experimental results demonstrate that C‐SSD outperforms other networks at three different image scales and achieves the best performance of 83. A 8% accuracy on the PASCAL VOC2007 test set, proving the effectiveness of the algorithm. C‐SSD achieves a better balance of speed and accuracy, showing excellent performance in rapid detection of small targets. [ABSTRACT FROM AUTHOR]

The blood-brain barrier (BBB) is a dynamic regulatory barrier at the interface of blood circulation and the brain parenchyma, which plays a critical role in protecting homeostasis in the central nervous system. However, it also significantly impedes drug delivery to the brain. Understanding the transport across BBB and brain distribution will facilitate the prediction of drug delivery efficiency and the development of new therapies. To date, various methods and models have been developed to study drug transport at the BBB interface, including in vivo brain uptake measurement methods, in vitro BBB models, and mathematic brain vascular models. Since the in vitro BBB models have been extensively reviewed elsewhere, we provide a comprehensive summary of the brain transport mechanisms and the currently available in vivo methods and mathematic models in studying the molecule delivery process at the BBB interface. In particular, we reviewed the emerging in vivo imaging techniques in observing drug transport across the BBB. We discussed the advantages and disadvantages associated with each model to serve as a guide for model selection in studying drug transport across the BBB. In summary, we envision future directions to improve the accuracy of mathematical models, establish noninvasive in vivo measurement techniques, and bridge the preclinical studies with clinical translation by taking the altered BBB physiological conditions into consideration. We believe these are critical in guiding new drug development and precise drug administration in brain disease treatment., (Copyright © 2023 by ASME.)