近年来,基于深度学习的低光图像增强方法受Retinex理论的启发,先估计照明图调整亮度,再恢复反射率以实现低光增强。因此,通过分析低光场景反射图与参考反射图的相似度,提出一种反射先验图引导的低光图像增强网络(RP-Net)。首先,在Lab色...近年来,基于深度学习的低光图像增强方法受Retinex理论的启发,先估计照明图调整亮度,再恢复反射率以实现低光增强。因此,通过分析低光场景反射图与参考反射图的相似度,提出一种反射先验图引导的低光图像增强网络(RP-Net)。首先,在Lab色彩空间分解出相似反射图,并设计反射先验特征自适应提取器(RPAE)在主干网络以不同尺度从相似反射图中重新编码和筛选引导特征;其次,通过设计的反射先验特征引导注意力块(RPGB)将引导信息注入主干网络。此外,针对传统逐像素L1损失的局限性,从频域分析的视角出发,设计一种频域调和损失函数,以从全局频谱分布优化增强效果。在LOLv1、LOLv2和LSRW数据集上的实验结果表明,所提方法在结构相似性(SSIM)上优于现有主流方法,在LOLv2-syn和LSRW数据集上得到的峰值信噪比(PSNR)相较于Retinexformer和SAFNet(Spatial And Frequency Network)分别提高了1.29 dB和2.08 dB,并且在色彩保真和增强效果的平衡上表现出色。展开更多
BACKGROUND Orthopaedic surgical education has traditionally depended on the apprenticeship model of“see one,do one,teach one”.However,reduced operative exposure,stricter work-hour regulations,medicolegal constraints...BACKGROUND Orthopaedic surgical education has traditionally depended on the apprenticeship model of“see one,do one,teach one”.However,reduced operative exposure,stricter work-hour regulations,medicolegal constraints,and patient safety concerns have constrained its practicality.Simulation-based training has become a reliable,safe,and cost-efficient alternative.Dry lab techniques,especially virtual and augmented reality,make up 78%of current dry lab research,whereas wet labs still set the standard for anatomical realism.AIM To evaluate the effectiveness,limitations,and future directions of wet and dry lab simulation in orthopaedic training.METHODS A scoping review was carried out across four databases-PubMed,Cochrane Library,Web of Science,and EBSCOhost-up to 2025.Medical Subject Headings included:"Orthopaedic Education","Wet Lab","Dry Lab","Simulation Training","Virtual Reality",and"Surgical Procedure".Eligible studies focused on orthopaedic or spinal surgical education,employed wet or dry lab techniques,and assessed training effectiveness.Exclusion criteria consisted of non-English publications,abstracts only,non-orthopaedic research,and studies unrelated to simulation.Two reviewers independently screened titles,abstracts,and full texts,resolving discrepancies with a third reviewer.RESULTS From 1851 records,101 studies met inclusion:78 on dry labs,7 on wet labs,4 on both.Virtual reality(VR)simulations were most common,with AI increasingly used for feedback and assessment.Cadaveric training remains the gold standard for accuracy and tactile feedback,while dry labs-especially VR-offer scalability,lower cost(40%-60%savings in five studies),and accessibility for novices.Senior residents prefer wet labs for complex tasks;juniors favour dry labs for basics.Challenges include limited transferability data,lack of standard outcome metrics,and ethical concerns about cadaver use and AI assessment.CONCLUSION Wet and dry labs each have unique strengths in orthopaedic training.A hybrid approach combining both,supported by standardised assessments and outcome studies,is most effective.Future efforts should aim for uniform reporting,integrating new technologies,and policy support for hybrid curricula to enhance skills and patient care.展开更多
数字化赋能教学已成为当前教学的新趋势,本文借助Earth Space Lab程序赋能“地球的运动”教学,帮助学生从本质上理解较为抽象的地理概念,同时,基于Earth Space Lab的教学不仅能实现学生对抽象概念的熟练掌握与灵活应用,促进作业质量与...数字化赋能教学已成为当前教学的新趋势,本文借助Earth Space Lab程序赋能“地球的运动”教学,帮助学生从本质上理解较为抽象的地理概念,同时,基于Earth Space Lab的教学不仅能实现学生对抽象概念的熟练掌握与灵活应用,促进作业质量与学习效果显著提升,还能进一步提升学生学习兴趣、地理实践力与综合思维,促进其地理核心素养的培育。展开更多
文摘近年来,基于深度学习的低光图像增强方法受Retinex理论的启发,先估计照明图调整亮度,再恢复反射率以实现低光增强。因此,通过分析低光场景反射图与参考反射图的相似度,提出一种反射先验图引导的低光图像增强网络(RP-Net)。首先,在Lab色彩空间分解出相似反射图,并设计反射先验特征自适应提取器(RPAE)在主干网络以不同尺度从相似反射图中重新编码和筛选引导特征;其次,通过设计的反射先验特征引导注意力块(RPGB)将引导信息注入主干网络。此外,针对传统逐像素L1损失的局限性,从频域分析的视角出发,设计一种频域调和损失函数,以从全局频谱分布优化增强效果。在LOLv1、LOLv2和LSRW数据集上的实验结果表明,所提方法在结构相似性(SSIM)上优于现有主流方法,在LOLv2-syn和LSRW数据集上得到的峰值信噪比(PSNR)相较于Retinexformer和SAFNet(Spatial And Frequency Network)分别提高了1.29 dB和2.08 dB,并且在色彩保真和增强效果的平衡上表现出色。
文摘BACKGROUND Orthopaedic surgical education has traditionally depended on the apprenticeship model of“see one,do one,teach one”.However,reduced operative exposure,stricter work-hour regulations,medicolegal constraints,and patient safety concerns have constrained its practicality.Simulation-based training has become a reliable,safe,and cost-efficient alternative.Dry lab techniques,especially virtual and augmented reality,make up 78%of current dry lab research,whereas wet labs still set the standard for anatomical realism.AIM To evaluate the effectiveness,limitations,and future directions of wet and dry lab simulation in orthopaedic training.METHODS A scoping review was carried out across four databases-PubMed,Cochrane Library,Web of Science,and EBSCOhost-up to 2025.Medical Subject Headings included:"Orthopaedic Education","Wet Lab","Dry Lab","Simulation Training","Virtual Reality",and"Surgical Procedure".Eligible studies focused on orthopaedic or spinal surgical education,employed wet or dry lab techniques,and assessed training effectiveness.Exclusion criteria consisted of non-English publications,abstracts only,non-orthopaedic research,and studies unrelated to simulation.Two reviewers independently screened titles,abstracts,and full texts,resolving discrepancies with a third reviewer.RESULTS From 1851 records,101 studies met inclusion:78 on dry labs,7 on wet labs,4 on both.Virtual reality(VR)simulations were most common,with AI increasingly used for feedback and assessment.Cadaveric training remains the gold standard for accuracy and tactile feedback,while dry labs-especially VR-offer scalability,lower cost(40%-60%savings in five studies),and accessibility for novices.Senior residents prefer wet labs for complex tasks;juniors favour dry labs for basics.Challenges include limited transferability data,lack of standard outcome metrics,and ethical concerns about cadaver use and AI assessment.CONCLUSION Wet and dry labs each have unique strengths in orthopaedic training.A hybrid approach combining both,supported by standardised assessments and outcome studies,is most effective.Future efforts should aim for uniform reporting,integrating new technologies,and policy support for hybrid curricula to enhance skills and patient care.
文摘数字化赋能教学已成为当前教学的新趋势,本文借助Earth Space Lab程序赋能“地球的运动”教学,帮助学生从本质上理解较为抽象的地理概念,同时,基于Earth Space Lab的教学不仅能实现学生对抽象概念的熟练掌握与灵活应用,促进作业质量与学习效果显著提升,还能进一步提升学生学习兴趣、地理实践力与综合思维,促进其地理核心素养的培育。
