AIM:To compare visual field defects using the Swedish Interactive Thresholding Algorithm(SITA)Fast strategy with SITA Faster strategy,a newly developed time-saving threshold visual field strategy.METHODS:Ninety-three ...AIM:To compare visual field defects using the Swedish Interactive Thresholding Algorithm(SITA)Fast strategy with SITA Faster strategy,a newly developed time-saving threshold visual field strategy.METHODS:Ninety-three participants(60 glaucoma patients and 33 normal controls)were enrolled.One eye from each participant was selected randomly for the study.SITA Fast and SITA Faster were performed using the 24-2 default mode for each test.The differences of visual field defects between the two strategies were compared using the test duration,false-positive response errors,mean deviation(MD),visual field index(VFI)and the numbers of depressed test points at the significant levels of P<5%,<2%,<1%,and<0.5%in probability plots.The correlation between strategies was analyzed.The agreement between strategies was acquired by Bland-Altman analysis.RESULTS:Mean test durations were 246.0±60.9 s for SITA Fast,and 156.3±46.3 s for SITA Faster(P<0.001).The test duration of SITA Faster was 36.5%shorter than SITA Fast.The MD,VFI and numbers of depressed points at P<5%,<2%,<1%,and<0.5%in probability plots showed no statistically significant difference between two strategies(P>0.05).Correlation analysis showed a high correlation for MD(r=0.986,P<0.001)and VFI(r=0.986,P<0.001)between the two strategies.Bland-Altman analysis showed great agreement between the two strategies.CONCLUSION:SITA Faster,which saves considerable test time,has a great test quality comparing to SITA Fast,but may be not directly interchangeable.展开更多
With the continuous development of deep learning,Deep Convolutional Neural Network(DCNN)has attracted wide attention in the industry due to its high accuracy in image classification.Compared with other DCNN hard-ware ...With the continuous development of deep learning,Deep Convolutional Neural Network(DCNN)has attracted wide attention in the industry due to its high accuracy in image classification.Compared with other DCNN hard-ware deployment platforms,Field Programmable Gate Array(FPGA)has the advantages of being programmable,low power consumption,parallelism,and low cost.However,the enormous amount of calculation of DCNN and the limited logic capacity of FPGA restrict the energy efficiency of the DCNN accelerator.The traditional sequential sliding window method can improve the throughput of the DCNN accelerator by data multiplexing,but this method’s data multiplexing rate is low because it repeatedly reads the data between rows.This paper proposes a fast data readout strategy via the circular sliding window data reading method,it can improve the multiplexing rate of data between rows by optimizing the memory access order of input data.In addition,the multiplication bit width of the DCNN accelerator is much smaller than that of the Digital Signal Processing(DSP)on the FPGA,which means that there will be a waste of resources if a multiplication uses a single DSP.A multiplier sharing strategy is proposed,the multiplier of the accelerator is customized so that a single DSP block can complete multiple groups of 4,6,and 8-bit signed multiplication in parallel.Finally,based on two strategies of appeal,an FPGA optimized accelerator is proposed.The accelerator is customized by Verilog language and deployed on Xilinx VCU118.When the accelerator recognizes the CIRFAR-10 dataset,its energy efficiency is 39.98 GOPS/W,which provides 1.73×speedup energy efficiency over previous DCNN FPGA accelerators.When the accelerator recognizes the IMAGENET dataset,its energy efficiency is 41.12 GOPS/W,which shows 1.28×−3.14×energy efficiency compared with others.展开更多
To reduce the carbon footprint in the transportation sector and improve overall vehicle efficiency,a large number of electric vehicles are being manufactured.This is due to the fact that environmental concerns and the...To reduce the carbon footprint in the transportation sector and improve overall vehicle efficiency,a large number of electric vehicles are being manufactured.This is due to the fact that environmental concerns and the depletion of fossil fuels have become significant global problems.Lithium-ion batteries(LIBs)have been distinguished themselves from alternative energy storage technologies for electric vehicles(EVs) due to superior qualities like high energy and power density,extended cycle life,and low maintenance cost to a competitive price.However,there are still certain challenges to be solved,like EV fast charging,longer lifetime,and reduced weight.For fast charging,the multi-stage constant current(MSCC) charging technique is an emerging solution to improve charging efficiency,reduce temperature rise during charging,increase charging/discharging capacities,shorten charging time,and extend the cycle life.However,there are large variations in the implementation of the number of stages,stage transition criterion,and C-rate selection for each stage.This paper provides a review of these problems by compiling information from the literature.An overview of the impact of different design parameters(number of stages,stage transition,and C-rate) that the MSCC charging techniques have had on the LIB performance and cycle life is described in detail and analyzed.The impact of design parameters on lifetime,charging efficiency,charging and discharging capacity,charging speed,and rising temperature during charging is presented,and this review provides guidelines for designing advanced fast charging strategies and determining future research gaps.展开更多
文摘AIM:To compare visual field defects using the Swedish Interactive Thresholding Algorithm(SITA)Fast strategy with SITA Faster strategy,a newly developed time-saving threshold visual field strategy.METHODS:Ninety-three participants(60 glaucoma patients and 33 normal controls)were enrolled.One eye from each participant was selected randomly for the study.SITA Fast and SITA Faster were performed using the 24-2 default mode for each test.The differences of visual field defects between the two strategies were compared using the test duration,false-positive response errors,mean deviation(MD),visual field index(VFI)and the numbers of depressed test points at the significant levels of P<5%,<2%,<1%,and<0.5%in probability plots.The correlation between strategies was analyzed.The agreement between strategies was acquired by Bland-Altman analysis.RESULTS:Mean test durations were 246.0±60.9 s for SITA Fast,and 156.3±46.3 s for SITA Faster(P<0.001).The test duration of SITA Faster was 36.5%shorter than SITA Fast.The MD,VFI and numbers of depressed points at P<5%,<2%,<1%,and<0.5%in probability plots showed no statistically significant difference between two strategies(P>0.05).Correlation analysis showed a high correlation for MD(r=0.986,P<0.001)and VFI(r=0.986,P<0.001)between the two strategies.Bland-Altman analysis showed great agreement between the two strategies.CONCLUSION:SITA Faster,which saves considerable test time,has a great test quality comparing to SITA Fast,but may be not directly interchangeable.
基金supported in part by the Major Program of the Ministry of Science and Technology of China under Grant 2019YFB2205102in part by the National Natural Science Foundation of China under Grant 61974164,62074166,61804181,62004219,62004220,62104256.
文摘With the continuous development of deep learning,Deep Convolutional Neural Network(DCNN)has attracted wide attention in the industry due to its high accuracy in image classification.Compared with other DCNN hard-ware deployment platforms,Field Programmable Gate Array(FPGA)has the advantages of being programmable,low power consumption,parallelism,and low cost.However,the enormous amount of calculation of DCNN and the limited logic capacity of FPGA restrict the energy efficiency of the DCNN accelerator.The traditional sequential sliding window method can improve the throughput of the DCNN accelerator by data multiplexing,but this method’s data multiplexing rate is low because it repeatedly reads the data between rows.This paper proposes a fast data readout strategy via the circular sliding window data reading method,it can improve the multiplexing rate of data between rows by optimizing the memory access order of input data.In addition,the multiplication bit width of the DCNN accelerator is much smaller than that of the Digital Signal Processing(DSP)on the FPGA,which means that there will be a waste of resources if a multiplication uses a single DSP.A multiplier sharing strategy is proposed,the multiplier of the accelerator is customized so that a single DSP block can complete multiple groups of 4,6,and 8-bit signed multiplication in parallel.Finally,based on two strategies of appeal,an FPGA optimized accelerator is proposed.The accelerator is customized by Verilog language and deployed on Xilinx VCU118.When the accelerator recognizes the CIRFAR-10 dataset,its energy efficiency is 39.98 GOPS/W,which provides 1.73×speedup energy efficiency over previous DCNN FPGA accelerators.When the accelerator recognizes the IMAGENET dataset,its energy efficiency is 41.12 GOPS/W,which shows 1.28×−3.14×energy efficiency compared with others.
文摘To reduce the carbon footprint in the transportation sector and improve overall vehicle efficiency,a large number of electric vehicles are being manufactured.This is due to the fact that environmental concerns and the depletion of fossil fuels have become significant global problems.Lithium-ion batteries(LIBs)have been distinguished themselves from alternative energy storage technologies for electric vehicles(EVs) due to superior qualities like high energy and power density,extended cycle life,and low maintenance cost to a competitive price.However,there are still certain challenges to be solved,like EV fast charging,longer lifetime,and reduced weight.For fast charging,the multi-stage constant current(MSCC) charging technique is an emerging solution to improve charging efficiency,reduce temperature rise during charging,increase charging/discharging capacities,shorten charging time,and extend the cycle life.However,there are large variations in the implementation of the number of stages,stage transition criterion,and C-rate selection for each stage.This paper provides a review of these problems by compiling information from the literature.An overview of the impact of different design parameters(number of stages,stage transition,and C-rate) that the MSCC charging techniques have had on the LIB performance and cycle life is described in detail and analyzed.The impact of design parameters on lifetime,charging efficiency,charging and discharging capacity,charging speed,and rising temperature during charging is presented,and this review provides guidelines for designing advanced fast charging strategies and determining future research gaps.
