Visual perception computing is a crucial area in artificial intelligence,aiming to simulate human vision for the intelligent analysis of complex visual data.However,current methods face several challenges,such as miss...Visual perception computing is a crucial area in artificial intelligence,aiming to simulate human vision for the intelligent analysis of complex visual data.However,current methods face several challenges,such as missing data,weak generalization across different scenarios,and difficulties in learning complex patterns,particularly in rare or long-tail situations.The framework of parallel images is reviewed in this paper,which provides new ways to advance visual perception by closely connecting real imaging systems with artificial ones.First,artificial image systems can be built to reflect real environments,enabling both real and artificial images to work together.These artificial systems produce multi-modal data,helping to solve the problem of incomplete data.Second,virtual-to-real model transfer approaches based on multi-view feature fusion are discussed,which support adaptive model improvement and better generalization to new scenarios.Finally,parallel visual models are introduced that combine data from different sources and integrate various types of knowledge,greatly improving performance on diverse visual recognition tasks.展开更多
Self-driving and recreational vehicle(RV)camps are a new form of industry module with the integration of transportation and tourism in China,thus the scientific and reasonable site selection plays an important role in...Self-driving and recreational vehicle(RV)camps are a new form of industry module with the integration of transportation and tourism in China,thus the scientific and reasonable site selection plays an important role in the success of camps’construction and operation.In terms of relying resources and development factors,camps can be divided into five categories:scenic-spot-based,transportation-based,environment-based,project-based and leisure and vacation-based.According to whether it is of excludability and competitiveness,the camps in China mainly embody the attribute of private products.Based on the combination of subjective evaluation and objective calculation,the evaluation model of spatial site selection is constructed and the weight of each index is calculated by using analytic hierarchy process(AHP)method and entropy coefficient method.The results show that traffic condition factor is the priority to the selection of campsite,and whether it is on the popular main self-driving route and the grade of trunk roads are the dominant indices.The second factor taken into consideration is the social factors,in which government policy supports and land cost play a key role.The third factor is the market,in which the urban economic level,partnership with the government and tourist resource conditionsare of great importance.The fourth factor of the campsite selection includes natural elements,in which the quality of ecological environment and water source conditions are mostly considered.In the future,it is suggested that a camp pattern of"public goods"plus"private goods"should be built and the construction of camps in underdeveloped areas should be highly developed so as to form camp spatial network from individual points to a series of campsite and finally the campsite group in China will be set up.展开更多
Background Bronchopulmonary dysplasia(BPD)is a common chronic lung disease in extremely preterm neonates.The outcome and clinical burden vary dramatically according to severity.Although some prediction tools for BPD e...Background Bronchopulmonary dysplasia(BPD)is a common chronic lung disease in extremely preterm neonates.The outcome and clinical burden vary dramatically according to severity.Although some prediction tools for BPD exist,they seldom pay attention to disease severity and are based on populations in developed countries.This study aimed to develop machine learning prediction models for BPD severity based on selected clinical factors in a Chinese population.Methods In this retrospective,single-center study,we included patients with a gestational age<32 weeks who were diagnosed with BPD in our neonatal intensive care unit from 2016 to 2020.We collected their clinical information during the maternal,birth and early postnatal periods.Risk factors were selected through univariable and ordinal logistic regression analyses.Prediction models based on logistic regression(LR),gradient boosting decision tree,XGBoost(XGB)and random forest(RF)models were implemented and assessed by the area under the receiver operating characteristic curve(AUC).Results We ultimately included 471 patients(279 mild,147 moderate,and 45 severe cases).On ordinal logistic regression,gestational diabetes mellitus,initial fraction of inspiration O_(2) value,invasive ventilation,acidosis,hypochloremia,C-reactive protein level,patent ductus arteriosus and Gram-negative respiratory culture were independent risk factors for BPD severity.All the XGB,LR and RF models(AUC=0.85,0.86 and 0.84,respectively)all had good performance.Conclusions We found risk factors for BPD severity in our population and developed machine learning models based on them.The models have good performance and can be used to aid in predicting BPD severity in the Chinese population.展开更多
基金supported by the Natural Science Foundation of China(Nos.62203040 and 62472048)the Beijing Natural Science Foundation(No.L242081).
文摘Visual perception computing is a crucial area in artificial intelligence,aiming to simulate human vision for the intelligent analysis of complex visual data.However,current methods face several challenges,such as missing data,weak generalization across different scenarios,and difficulties in learning complex patterns,particularly in rare or long-tail situations.The framework of parallel images is reviewed in this paper,which provides new ways to advance visual perception by closely connecting real imaging systems with artificial ones.First,artificial image systems can be built to reflect real environments,enabling both real and artificial images to work together.These artificial systems produce multi-modal data,helping to solve the problem of incomplete data.Second,virtual-to-real model transfer approaches based on multi-view feature fusion are discussed,which support adaptive model improvement and better generalization to new scenarios.Finally,parallel visual models are introduced that combine data from different sources and integrate various types of knowledge,greatly improving performance on diverse visual recognition tasks.
基金sponsored by the National Social Science Fund of China(grant number 20&ZD099)with the project name“research on the spatial effects of China’s cross regional major infrastructure”。
文摘Self-driving and recreational vehicle(RV)camps are a new form of industry module with the integration of transportation and tourism in China,thus the scientific and reasonable site selection plays an important role in the success of camps’construction and operation.In terms of relying resources and development factors,camps can be divided into five categories:scenic-spot-based,transportation-based,environment-based,project-based and leisure and vacation-based.According to whether it is of excludability and competitiveness,the camps in China mainly embody the attribute of private products.Based on the combination of subjective evaluation and objective calculation,the evaluation model of spatial site selection is constructed and the weight of each index is calculated by using analytic hierarchy process(AHP)method and entropy coefficient method.The results show that traffic condition factor is the priority to the selection of campsite,and whether it is on the popular main self-driving route and the grade of trunk roads are the dominant indices.The second factor taken into consideration is the social factors,in which government policy supports and land cost play a key role.The third factor is the market,in which the urban economic level,partnership with the government and tourist resource conditionsare of great importance.The fourth factor of the campsite selection includes natural elements,in which the quality of ecological environment and water source conditions are mostly considered.In the future,it is suggested that a camp pattern of"public goods"plus"private goods"should be built and the construction of camps in underdeveloped areas should be highly developed so as to form camp spatial network from individual points to a series of campsite and finally the campsite group in China will be set up.
基金funded by the Science and Technology Commission of Shanghai Municipality(No.21511104502)the National Key Research and Development Program of China(No.2021ZD0113501).
文摘Background Bronchopulmonary dysplasia(BPD)is a common chronic lung disease in extremely preterm neonates.The outcome and clinical burden vary dramatically according to severity.Although some prediction tools for BPD exist,they seldom pay attention to disease severity and are based on populations in developed countries.This study aimed to develop machine learning prediction models for BPD severity based on selected clinical factors in a Chinese population.Methods In this retrospective,single-center study,we included patients with a gestational age<32 weeks who were diagnosed with BPD in our neonatal intensive care unit from 2016 to 2020.We collected their clinical information during the maternal,birth and early postnatal periods.Risk factors were selected through univariable and ordinal logistic regression analyses.Prediction models based on logistic regression(LR),gradient boosting decision tree,XGBoost(XGB)and random forest(RF)models were implemented and assessed by the area under the receiver operating characteristic curve(AUC).Results We ultimately included 471 patients(279 mild,147 moderate,and 45 severe cases).On ordinal logistic regression,gestational diabetes mellitus,initial fraction of inspiration O_(2) value,invasive ventilation,acidosis,hypochloremia,C-reactive protein level,patent ductus arteriosus and Gram-negative respiratory culture were independent risk factors for BPD severity.All the XGB,LR and RF models(AUC=0.85,0.86 and 0.84,respectively)all had good performance.Conclusions We found risk factors for BPD severity in our population and developed machine learning models based on them.The models have good performance and can be used to aid in predicting BPD severity in the Chinese population.
