目的构建融合分区注意力的UNet(PA-UNet)模型,观察其分割MRI中的膝关节软骨的价值。方法对来源于Osteoarthritis Initiative-Zuse Institute Berlin数据集的膝关节MRI进行切片及预处理,以UNet为骨干网络构建基于分区注意力机制的PA-UNe...目的构建融合分区注意力的UNet(PA-UNet)模型,观察其分割MRI中的膝关节软骨的价值。方法对来源于Osteoarthritis Initiative-Zuse Institute Berlin数据集的膝关节MRI进行切片及预处理,以UNet为骨干网络构建基于分区注意力机制的PA-UNet模型,通过主、客观评价比较该模型与其他模型分割股骨软骨及胫骨软骨的效果;分别以基于UNet、基于SE(第2~4层)的UNet(UNet+SE)、+UNet、++UNet、+++UNet、+UNet+、++UNet++及PA-UNet模型的消融实验观察各模型分割膝关节软骨的效果。结果PA-UNet可准确分割低难度、中等难度及困难样本中的股骨及胫骨软骨,其分割细小结构效果优于SegNet、UNet及DeepLabv3+模型;其分割股骨软骨及胫骨软骨的戴斯相似系数(DSC)及交并比均高于、而豪斯多夫距离均低于UNet、DeepLabv3+、SA-UNet、RA UNet及SegNet模型。以PA-UNet模型分割股骨软骨及胫骨软骨的DSC分别为88.97%及82.72%,均高于UNet、UNet+SE、+UNet、++UNet、+++UNet、+UNet+及++UNet++模型。结论PA-UNet可完整分割MRI中的膝关节软骨,尤其对细小结构的分割效果良好。展开更多
Semantic segmentation is a crucial step for document understanding.In this paper,an NVIDIA Jetson Nano-based platform is applied for implementing semantic segmentation for teaching artificial intelligence concepts and...Semantic segmentation is a crucial step for document understanding.In this paper,an NVIDIA Jetson Nano-based platform is applied for implementing semantic segmentation for teaching artificial intelligence concepts and programming.To extract semantic structures from document images,we present an end-to-end dilated convolution network architecture.Dilated convolutions have well-known advantages for extracting multi-scale context information without losing spatial resolution.Our model utilizes dilated convolutions with residual network to represent the image features and predicting pixel labels.The convolution part works as feature extractor to obtain multidimensional and hierarchical image features.The consecutive deconvolution is used for producing full resolution segmentation prediction.The probability of each pixel decides its predefined semantic class label.To understand segmentation granularity,we compare performances at three different levels.From fine grained class to coarse class levels,the proposed dilated convolution network architecture is evaluated on three document datasets.The experimental results have shown that both semantic data distribution imbalance and network depth are import factors that influence the document’s semantic segmentation performances.The research is aimed at offering an education resource for teaching artificial intelligence concepts and techniques.展开更多
The development of practical methods for the direct and selective C(sp^(3))‒H functionalization of hydrocarbons is an attractive topic in synthetic chemistry.Although the radical‐mediated hydrogen atom transfer(HAT)p...The development of practical methods for the direct and selective C(sp^(3))‒H functionalization of hydrocarbons is an attractive topic in synthetic chemistry.Although the radical‐mediated hydrogen atom transfer(HAT)process has shown considerable potential in such reactions,it still faces fundamental problems associated with reactivity and selectivity.Herein,we report a convenient and economic approach to site‐selective C(sp^(3))‒H sulfonylation via photo‐induced HAT catalysis.Employing a conjugated polycyclic quinone as a direct HAT photocatalyst,commercially available inorganic sulfinates as the sulfonylation source,copper triflate as an inexpensive oxidant,a variety of toluene derivatives and cycloalkanes were converted into biologically and synthetically interesting sulfone products under mild conditions.The mechanistic studies reveal that the reaction sequence involves direct HAT‐induced radical formation and a subsequent copper‐mediated organometallic process for the C‒S bond formation.This method offers an appealing opportunity to furnish high value‐added products from abundant hydrocarbon starting materials and inexpensive reagents.展开更多
Regiodivergent asymmetric cycloadditions from the same set of starting materials offer interesting opportunities for rapid construction of optically active cyclic molecules with structural diversity.However,this remai...Regiodivergent asymmetric cycloadditions from the same set of starting materials offer interesting opportunities for rapid construction of optically active cyclic molecules with structural diversity.However,this remains a challenging task due to the difficulty of simultaneously controlling the regio-,diastereo-,and enantioselectivity in the ring formation processes.To address this long-standing problem.展开更多
文摘目的构建融合分区注意力的UNet(PA-UNet)模型,观察其分割MRI中的膝关节软骨的价值。方法对来源于Osteoarthritis Initiative-Zuse Institute Berlin数据集的膝关节MRI进行切片及预处理,以UNet为骨干网络构建基于分区注意力机制的PA-UNet模型,通过主、客观评价比较该模型与其他模型分割股骨软骨及胫骨软骨的效果;分别以基于UNet、基于SE(第2~4层)的UNet(UNet+SE)、+UNet、++UNet、+++UNet、+UNet+、++UNet++及PA-UNet模型的消融实验观察各模型分割膝关节软骨的效果。结果PA-UNet可准确分割低难度、中等难度及困难样本中的股骨及胫骨软骨,其分割细小结构效果优于SegNet、UNet及DeepLabv3+模型;其分割股骨软骨及胫骨软骨的戴斯相似系数(DSC)及交并比均高于、而豪斯多夫距离均低于UNet、DeepLabv3+、SA-UNet、RA UNet及SegNet模型。以PA-UNet模型分割股骨软骨及胫骨软骨的DSC分别为88.97%及82.72%,均高于UNet、UNet+SE、+UNet、++UNet、+++UNet、+UNet+及++UNet++模型。结论PA-UNet可完整分割MRI中的膝关节软骨,尤其对细小结构的分割效果良好。
基金Project(61806107)supported by the National Natural Science Foundation of ChinaProject supported by the Shandong Key Laboratory of Wisdom Mine Information Technology,ChinaProject supported by the Opening Project of State Key Laboratory of Digital Publishing Technology,China。
文摘Semantic segmentation is a crucial step for document understanding.In this paper,an NVIDIA Jetson Nano-based platform is applied for implementing semantic segmentation for teaching artificial intelligence concepts and programming.To extract semantic structures from document images,we present an end-to-end dilated convolution network architecture.Dilated convolutions have well-known advantages for extracting multi-scale context information without losing spatial resolution.Our model utilizes dilated convolutions with residual network to represent the image features and predicting pixel labels.The convolution part works as feature extractor to obtain multidimensional and hierarchical image features.The consecutive deconvolution is used for producing full resolution segmentation prediction.The probability of each pixel decides its predefined semantic class label.To understand segmentation granularity,we compare performances at three different levels.From fine grained class to coarse class levels,the proposed dilated convolution network architecture is evaluated on three document datasets.The experimental results have shown that both semantic data distribution imbalance and network depth are import factors that influence the document’s semantic segmentation performances.The research is aimed at offering an education resource for teaching artificial intelligence concepts and techniques.
文摘The development of practical methods for the direct and selective C(sp^(3))‒H functionalization of hydrocarbons is an attractive topic in synthetic chemistry.Although the radical‐mediated hydrogen atom transfer(HAT)process has shown considerable potential in such reactions,it still faces fundamental problems associated with reactivity and selectivity.Herein,we report a convenient and economic approach to site‐selective C(sp^(3))‒H sulfonylation via photo‐induced HAT catalysis.Employing a conjugated polycyclic quinone as a direct HAT photocatalyst,commercially available inorganic sulfinates as the sulfonylation source,copper triflate as an inexpensive oxidant,a variety of toluene derivatives and cycloalkanes were converted into biologically and synthetically interesting sulfone products under mild conditions.The mechanistic studies reveal that the reaction sequence involves direct HAT‐induced radical formation and a subsequent copper‐mediated organometallic process for the C‒S bond formation.This method offers an appealing opportunity to furnish high value‐added products from abundant hydrocarbon starting materials and inexpensive reagents.
基金from the National Natural Science Foundation of China(grant nos.22071209 and 22071206)the National Youth Talent Support Program,the Natural Science Foundation of Fujian Province of China(grant no.2017J06006)+1 种基金the Fundamental Research Funds for the Central Universities(grant no.20720190048)Basic Disciplines Training Program for Top-notch Students of the Ministry of Education.
文摘Regiodivergent asymmetric cycloadditions from the same set of starting materials offer interesting opportunities for rapid construction of optically active cyclic molecules with structural diversity.However,this remains a challenging task due to the difficulty of simultaneously controlling the regio-,diastereo-,and enantioselectivity in the ring formation processes.To address this long-standing problem.
