文章提出一种在片上系统(System on Chip,SoC)实现高吞吐率的有限状态熵编码(finite state entropy,FSE)算法。通过压缩率、速度、资源消耗、功耗4个方面对所提出的编码器和解码器与典型的硬件哈夫曼编码(Huffman coding,HC)进行性能比...文章提出一种在片上系统(System on Chip,SoC)实现高吞吐率的有限状态熵编码(finite state entropy,FSE)算法。通过压缩率、速度、资源消耗、功耗4个方面对所提出的编码器和解码器与典型的硬件哈夫曼编码(Huffman coding,HC)进行性能比较,结果表明,所提出的硬件FSE编码器和解码器具有显著优势。硬件FSE(hFSE)架构实现在SoC的处理系统和可编程逻辑块(programmable logic,PL)上,通过高级可扩展接口(Advanced eXtensible Interface 4,AXI4)总线连接SoC的处理系统和可编程逻辑块。算法测试显示,FSE算法在非均匀数据分布和大数据量情况下,具有更好的压缩率。该文设计的编码器和解码器已在可编程逻辑块上实现,其中包括1个可配置的缓冲模块,将比特流作为单字节或双字节配置输出到8 bit位宽4096深度或16 bit位宽2048深度的块随机访问存储器(block random access memory,BRAM)中。所提出的FSE硬件架构为实时压缩应用提供了高吞吐率、低功耗和低资源消耗的硬件实现。展开更多
目的以关节镜检查结果为对照,比较三维稳态采集快速成像序列(3D FIESTA)和常规序列在诊断前交叉韧带(ACL)损伤中的准确性。资料与方法对50例患者的55个外伤膝关节行常规MRI(SET1WI,FSEPD/T2WI)和3D FIESTA序列扫描。依据MRI图像表现将AC...目的以关节镜检查结果为对照,比较三维稳态采集快速成像序列(3D FIESTA)和常规序列在诊断前交叉韧带(ACL)损伤中的准确性。资料与方法对50例患者的55个外伤膝关节行常规MRI(SET1WI,FSEPD/T2WI)和3D FIESTA序列扫描。依据MRI图像表现将ACL损伤分级,分析比较上述序列对ACL损伤诊断的敏感性、特异性、准确性和Kappa值。结果 3D FIESTA诊断ACL损伤的敏感性、特异性、准确性和Kappa值均较常规序列高,分别为94.6%、77.8%、89.1%、0.745。结论 3D FIESTA序列图像采集时间短,显示ACL损伤的准确率高,是常规膝关节序列检查的补充。展开更多
MULTIPLE sclerosis （MS） is a common inflammatory demyelinating disorder of central nervous system, and the disease burder could be well evaluated by conven-tional magnetic resonance imaging （MRI）,1 including T2-we...MULTIPLE sclerosis （MS） is a common inflammatory demyelinating disorder of central nervous system, and the disease burder could be well evaluated by conven-tional magnetic resonance imaging （MRI）,1 including T2-weighted, fluid-attenuatd inversion recovery, and postcontrast Tl-weighted sequences. 2 We investigated the perfusion state of MS plaques using brain perfusion imaging in a 12-year-old boy with MS.展开更多
With its rapid development in the wireless markets, IEEE 802.11 WLAN is experiencing a huge popularity. However, due to the limitation of frequency bandwidth of WLANs, it is essential that the available radio resource...With its rapid development in the wireless markets, IEEE 802.11 WLAN is experiencing a huge popularity. However, due to the limitation of frequency bandwidth of WLANs, it is essential that the available radio resource should be fully utilized to offer different services to multiple users. In order to maximize system throughput while still guaranteeing the fairness among users, a proportional fairness based algorithm is proposed in this work. Since most of the previous resource allocation algorithms were simply based on the channel conditions without taking into account user's demand, in this paper, we introduce the theory of fuzzy synthetic evaluation(FSE) which also allows us to consider user's demand as an important factor. As such, the fairness among users can be improved based on different users' requirements for services. In addition, a channel state information based rate adaptation scheme is also proposed. Through simulation studies, the results clearly validate that our proposed scheme shows advantages on providing user fairness while still improving the system throughput.展开更多
文摘文章提出一种在片上系统(System on Chip,SoC)实现高吞吐率的有限状态熵编码(finite state entropy,FSE)算法。通过压缩率、速度、资源消耗、功耗4个方面对所提出的编码器和解码器与典型的硬件哈夫曼编码(Huffman coding,HC)进行性能比较,结果表明,所提出的硬件FSE编码器和解码器具有显著优势。硬件FSE(hFSE)架构实现在SoC的处理系统和可编程逻辑块(programmable logic,PL)上,通过高级可扩展接口(Advanced eXtensible Interface 4,AXI4)总线连接SoC的处理系统和可编程逻辑块。算法测试显示,FSE算法在非均匀数据分布和大数据量情况下,具有更好的压缩率。该文设计的编码器和解码器已在可编程逻辑块上实现,其中包括1个可配置的缓冲模块,将比特流作为单字节或双字节配置输出到8 bit位宽4096深度或16 bit位宽2048深度的块随机访问存储器(block random access memory,BRAM)中。所提出的FSE硬件架构为实时压缩应用提供了高吞吐率、低功耗和低资源消耗的硬件实现。
文摘目的以关节镜检查结果为对照,比较三维稳态采集快速成像序列(3D FIESTA)和常规序列在诊断前交叉韧带(ACL)损伤中的准确性。资料与方法对50例患者的55个外伤膝关节行常规MRI(SET1WI,FSEPD/T2WI)和3D FIESTA序列扫描。依据MRI图像表现将ACL损伤分级,分析比较上述序列对ACL损伤诊断的敏感性、特异性、准确性和Kappa值。结果 3D FIESTA诊断ACL损伤的敏感性、特异性、准确性和Kappa值均较常规序列高,分别为94.6%、77.8%、89.1%、0.745。结论 3D FIESTA序列图像采集时间短,显示ACL损伤的准确率高,是常规膝关节序列检查的补充。
文摘MULTIPLE sclerosis （MS） is a common inflammatory demyelinating disorder of central nervous system, and the disease burder could be well evaluated by conven-tional magnetic resonance imaging （MRI）,1 including T2-weighted, fluid-attenuatd inversion recovery, and postcontrast Tl-weighted sequences. 2 We investigated the perfusion state of MS plaques using brain perfusion imaging in a 12-year-old boy with MS.
基金partially supported by the Academy of Finland (Decision No. 284748, 288473)
文摘With its rapid development in the wireless markets, IEEE 802.11 WLAN is experiencing a huge popularity. However, due to the limitation of frequency bandwidth of WLANs, it is essential that the available radio resource should be fully utilized to offer different services to multiple users. In order to maximize system throughput while still guaranteeing the fairness among users, a proportional fairness based algorithm is proposed in this work. Since most of the previous resource allocation algorithms were simply based on the channel conditions without taking into account user's demand, in this paper, we introduce the theory of fuzzy synthetic evaluation(FSE) which also allows us to consider user's demand as an important factor. As such, the fairness among users can be improved based on different users' requirements for services. In addition, a channel state information based rate adaptation scheme is also proposed. Through simulation studies, the results clearly validate that our proposed scheme shows advantages on providing user fairness while still improving the system throughput.
