While recent years have witnessed a dramatic upsurge of exploiting deep neural networks toward solving image denoising,existing methods mostly rely on simple noise assumptions,such as additive white Gaussian noise(AWG...While recent years have witnessed a dramatic upsurge of exploiting deep neural networks toward solving image denoising,existing methods mostly rely on simple noise assumptions,such as additive white Gaussian noise(AWGN),JPEG compression noise and camera sensor noise,and a general-purpose blind denoising method for real images remains unsolved.In this paper,we attempt to solve this problem from the perspective of network architecture design and training data synthesis.Specifically,for the network architecture design,we propose a swin-conv block to incorporate the local modeling ability of residual convolutional layer and non-local modeling ability of swin transformer block,and then plug it as the main building block into the widely-used image-to-image translation UNet architecture.For the training data synthesis,we design a practical noise degradation model which takes into consideration different kinds of noise(including Gaussian,Poisson,speckle,JPEG compression,and processed camera sensor noises)and resizing,and also involves a random shuffle strategy and a double degradation strategy.Extensive experiments on AGWN removal and real image denoising demonstrate that the new network architecture design achieves state-of-the-art performance and the new degradation model can help to significantly improve the practicability.We believe our work can provide useful insights into current denoising research.The source code is available at https://github.com/cszn/SCUNet.展开更多
Background:Single-cell RNA sequencing(scRNA-seq)technology is now becoming a widely applied method of transcriptome exploration that helps to reveal cell-type composition as well as cell-state heterogeneity for specif...Background:Single-cell RNA sequencing(scRNA-seq)technology is now becoming a widely applied method of transcriptome exploration that helps to reveal cell-type composition as well as cell-state heterogeneity for specific biological processes.Distinct sequencing platforms and processing pipelines may contribute to various results even for the same sequencing samples.Therefore,benchmarking sequencing platforms and processing pipelines was considered as a necessary step to interpret scRNA-seq data.However,recent comparing efforts were constrained in sequencing platforms or analyzing pipelines.There is still a lack of knowledge of analyzing pipelines matched with specific sequencing platforms in aspects of sensitivity,precision,and so on.Methods:We downloaded public scRNA-seq data that was generated by two distinct sequencers,NovaSeq 6000 and MGISEQ 2000.Then data was processed through the Drop-seq-tools,UMI-tools and Cell Ranger pipeline respectively.We calculated multiple measurements based on the expression profiles of the six platform-pipeline combinations.Results:We found that all three pipelines had comparable performance,the Cell Ranger pipeline achieved the best performance in precision while UMI-tools prevailed in terms of sensitivity and marker calling.Conclusions:Our work provided an insight into the selection of scRNA-seq data processing tools for two sequencing platforms as well as a framework to evaluate platform-pipeline combinations.展开更多
平方公里阵列(Square Kilometre Array,SKA)射电望远镜将在多个科学方向取得革命性的突破,而SKA软件系统是影响科学产品的关键因素之一.SKA区域中心是天文学家进行SKA数据分析、科学研究和学术交流的平台.处理SKA科学数据的软件环境需...平方公里阵列(Square Kilometre Array,SKA)射电望远镜将在多个科学方向取得革命性的突破,而SKA软件系统是影响科学产品的关键因素之一.SKA区域中心是天文学家进行SKA数据分析、科学研究和学术交流的平台.处理SKA科学数据的软件环境需要具备通用性、灵活性和高适应性.中国科学家已经建成了中国SKA区域中心原型机,部署了被大型超级计算机广泛使用的作业调度系统,并安装了能够处理当前主流射电望远镜观测数据的天文软件,还部署了多个科学数据处理管线,以方便不同科学方向的观测数据的自动化并行处理.本文介绍了中国SKA区域中心原型机的软件平台和处理SKA先导望远镜数据的管线,包括低频连续谱成像管线、谱线成像管线以及甚长基线干涉测量数据处理管线.国内外用户已经基于该平台成功开展了SKA相关科学研究.该平台的建设和运行为未来全面建设中国SKA区域中心提供了宝贵的实践经验.展开更多
基金This work was partly supported by the ETH Zürich Fund(OK),and by Huawei grants.
文摘While recent years have witnessed a dramatic upsurge of exploiting deep neural networks toward solving image denoising,existing methods mostly rely on simple noise assumptions,such as additive white Gaussian noise(AWGN),JPEG compression noise and camera sensor noise,and a general-purpose blind denoising method for real images remains unsolved.In this paper,we attempt to solve this problem from the perspective of network architecture design and training data synthesis.Specifically,for the network architecture design,we propose a swin-conv block to incorporate the local modeling ability of residual convolutional layer and non-local modeling ability of swin transformer block,and then plug it as the main building block into the widely-used image-to-image translation UNet architecture.For the training data synthesis,we design a practical noise degradation model which takes into consideration different kinds of noise(including Gaussian,Poisson,speckle,JPEG compression,and processed camera sensor noises)and resizing,and also involves a random shuffle strategy and a double degradation strategy.Extensive experiments on AGWN removal and real image denoising demonstrate that the new network architecture design achieves state-of-the-art performance and the new degradation model can help to significantly improve the practicability.We believe our work can provide useful insights into current denoising research.The source code is available at https://github.com/cszn/SCUNet.
基金This work was supported by Strategic Priority Research Program of Chinese Academy of Sciences(Nos.XDB38050200 and XDA26040304).
文摘Background:Single-cell RNA sequencing(scRNA-seq)technology is now becoming a widely applied method of transcriptome exploration that helps to reveal cell-type composition as well as cell-state heterogeneity for specific biological processes.Distinct sequencing platforms and processing pipelines may contribute to various results even for the same sequencing samples.Therefore,benchmarking sequencing platforms and processing pipelines was considered as a necessary step to interpret scRNA-seq data.However,recent comparing efforts were constrained in sequencing platforms or analyzing pipelines.There is still a lack of knowledge of analyzing pipelines matched with specific sequencing platforms in aspects of sensitivity,precision,and so on.Methods:We downloaded public scRNA-seq data that was generated by two distinct sequencers,NovaSeq 6000 and MGISEQ 2000.Then data was processed through the Drop-seq-tools,UMI-tools and Cell Ranger pipeline respectively.We calculated multiple measurements based on the expression profiles of the six platform-pipeline combinations.Results:We found that all three pipelines had comparable performance,the Cell Ranger pipeline achieved the best performance in precision while UMI-tools prevailed in terms of sensitivity and marker calling.Conclusions:Our work provided an insight into the selection of scRNA-seq data processing tools for two sequencing platforms as well as a framework to evaluate platform-pipeline combinations.
文摘平方公里阵列(Square Kilometre Array,SKA)射电望远镜将在多个科学方向取得革命性的突破,而SKA软件系统是影响科学产品的关键因素之一.SKA区域中心是天文学家进行SKA数据分析、科学研究和学术交流的平台.处理SKA科学数据的软件环境需要具备通用性、灵活性和高适应性.中国科学家已经建成了中国SKA区域中心原型机,部署了被大型超级计算机广泛使用的作业调度系统,并安装了能够处理当前主流射电望远镜观测数据的天文软件,还部署了多个科学数据处理管线,以方便不同科学方向的观测数据的自动化并行处理.本文介绍了中国SKA区域中心原型机的软件平台和处理SKA先导望远镜数据的管线,包括低频连续谱成像管线、谱线成像管线以及甚长基线干涉测量数据处理管线.国内外用户已经基于该平台成功开展了SKA相关科学研究.该平台的建设和运行为未来全面建设中国SKA区域中心提供了宝贵的实践经验.