基于传统暗原色先验原理的图像去雾算法存在的“halo”效应,且图像中明亮区域存在颜色失真现象,针对此问题,本文提出了多尺度窗口的自适应透射率修复交通图像去雾方法。首先,利用新的8方向边缘检测算子求取图像中景深突变区域,根据暗通...基于传统暗原色先验原理的图像去雾算法存在的“halo”效应,且图像中明亮区域存在颜色失真现象,针对此问题,本文提出了多尺度窗口的自适应透射率修复交通图像去雾方法。首先,利用新的8方向边缘检测算子求取图像中景深突变区域,根据暗通道先验理论和前一步求得的景深突变区域,在景深变化较大区域使用5 X 5的窗口,景深变化较小区域则使用15 x 15的窗口得到暗原色估计图。同时,针对暗通道先验原理对近景部分存在白色区域时透射率估计不准确的问题,引人了自适应透射率修复方法,通过引导滤波器得到边缘增强后的暗原色图像,并利用其与原暗原色图像的纹理差对近景区域的透射率进行修正,完成图像去雾。实验结果表明:双边滤波和梯度双边滤波两种算法均存在halo现象,并且在包含白色物体的明亮区域色彩失真严重,客观评价指标失去意义;相比于引导滤波,本文去雾算法的各项指标均有所提高,其中平均梯度平均提高了8.305%,PSNR平均提高了12.455%,边缘强度因子平均提高了7.77%。本文算法有效解决了复原图像中“halo”效应现象和明亮区域颜色失真现象,去雾效果最优。展开更多
Background： Prospective real-life data on the safety and effectiveness of rituximab in Chinese patients with diffuse large B-cell lymphoma （DLBCL） or follicular lymphoma （FL） are limited. This real-world study ai...Background： Prospective real-life data on the safety and effectiveness of rituximab in Chinese patients with diffuse large B-cell lymphoma （DLBCL） or follicular lymphoma （FL） are limited. This real-world study aimed to evaluate long-term safety and effectiveness outcomes ofrituximab plus chemotherapy （R-chemo） as first-line treatment in Chinese patients with DLBCL or FL. Hepatitis B virus （HBV） reactivation management was also investigated. Methods： A prospective, multicenter, single-arm, noninterventional study of previously untreated CD20-positive DLBCL or FL patients receiving first-line R-chemo treatment at 24 centers in China was conducted between January 17, 2011 and October 31, 2016. Enrolled patients underwent safety and effectiveness assessments after the last rituximab dose and were followed up for 3 years. Effectiveness endpoints included progression-free survival （PFS） and overall survival （OS）. Safety endpoints were adverse events （AEs）, serious AEs, drug-related AEs, and AEs of special interest. We also reported data on the incidence of HBV reactivation. Results： In total, 283 previously untreated CD20-positive DLBCL and 31 FL patients from 24 centers were enrolled. Three-year PFS was 59% （95% confidence interval [CI]： 50-67%） for DLBCL patients and 46% （95% CI： 20-69%） for FL patients. For DLBCL patients, multivariate analyses showed that PFS was not associated with international prognostic index, tumor maximum diameter, HBV infection status, or number ofrituximab treatment cycles, and OS was only associated with age 〉60 years （P 〈 0.05）. R-chemo was well tolerated. The incidence of HBV reactivation in hepatitis B surface antigen （HBsAg）-positive and HBsAg-negative/hepatitis B core antibody-positive patients was 13% （3/24） and 4% （3/69）, respectively. Conclusions： R-chemo is effective and safe in real-world clinical practice as first-line treatment for DLBCL and FL in China, and that HBV reactivation during R-chemo is manageable wi展开更多
研究RoboCup比赛未知环境中足球机器人的路径规划问题。提出一种多优化设计快速扩展随机树(rapidly exploring random tree,RRT)的足球机器人路径规划算法,解决了足球机器人在路径规划中存在的速度慢、效果差的问题。依据基本RRT算法原...研究RoboCup比赛未知环境中足球机器人的路径规划问题。提出一种多优化设计快速扩展随机树(rapidly exploring random tree,RRT)的足球机器人路径规划算法,解决了足球机器人在路径规划中存在的速度慢、效果差的问题。依据基本RRT算法原理,针对其随机性强、收敛速度慢以及路径平滑效果差的缺点,提出了随机采样点处增加引力分量、多步扩展逼近目标点以及冗余节点的剪裁与路径平滑等优化方式。在不同障碍物地图中的仿真实验表明,优化的规划路径长度比基本快速扩展随机树算法所得路径缩短约20%~30%,采样点数量减少45%~65%。最终将优化算法移植到SimRobot仿真平台,结果表明多优化设计RRT算法在未知环境中具备良好的实时性和鲁棒性,能够满足机器人在赛场上的路径规划要求。展开更多
文摘基于传统暗原色先验原理的图像去雾算法存在的“halo”效应,且图像中明亮区域存在颜色失真现象,针对此问题,本文提出了多尺度窗口的自适应透射率修复交通图像去雾方法。首先,利用新的8方向边缘检测算子求取图像中景深突变区域,根据暗通道先验理论和前一步求得的景深突变区域,在景深变化较大区域使用5 X 5的窗口,景深变化较小区域则使用15 x 15的窗口得到暗原色估计图。同时,针对暗通道先验原理对近景部分存在白色区域时透射率估计不准确的问题,引人了自适应透射率修复方法,通过引导滤波器得到边缘增强后的暗原色图像,并利用其与原暗原色图像的纹理差对近景区域的透射率进行修正,完成图像去雾。实验结果表明:双边滤波和梯度双边滤波两种算法均存在halo现象,并且在包含白色物体的明亮区域色彩失真严重,客观评价指标失去意义;相比于引导滤波,本文去雾算法的各项指标均有所提高,其中平均梯度平均提高了8.305%,PSNR平均提高了12.455%,边缘强度因子平均提高了7.77%。本文算法有效解决了复原图像中“halo”效应现象和明亮区域颜色失真现象,去雾效果最优。
基金This study was supported by grants from the National Natural Science Foundation of China （No. 81570186） and the Health and Family Planning Commission of Jiangsu Province （No. H201511）.
文摘Background： Prospective real-life data on the safety and effectiveness of rituximab in Chinese patients with diffuse large B-cell lymphoma （DLBCL） or follicular lymphoma （FL） are limited. This real-world study aimed to evaluate long-term safety and effectiveness outcomes ofrituximab plus chemotherapy （R-chemo） as first-line treatment in Chinese patients with DLBCL or FL. Hepatitis B virus （HBV） reactivation management was also investigated. Methods： A prospective, multicenter, single-arm, noninterventional study of previously untreated CD20-positive DLBCL or FL patients receiving first-line R-chemo treatment at 24 centers in China was conducted between January 17, 2011 and October 31, 2016. Enrolled patients underwent safety and effectiveness assessments after the last rituximab dose and were followed up for 3 years. Effectiveness endpoints included progression-free survival （PFS） and overall survival （OS）. Safety endpoints were adverse events （AEs）, serious AEs, drug-related AEs, and AEs of special interest. We also reported data on the incidence of HBV reactivation. Results： In total, 283 previously untreated CD20-positive DLBCL and 31 FL patients from 24 centers were enrolled. Three-year PFS was 59% （95% confidence interval [CI]： 50-67%） for DLBCL patients and 46% （95% CI： 20-69%） for FL patients. For DLBCL patients, multivariate analyses showed that PFS was not associated with international prognostic index, tumor maximum diameter, HBV infection status, or number ofrituximab treatment cycles, and OS was only associated with age 〉60 years （P 〈 0.05）. R-chemo was well tolerated. The incidence of HBV reactivation in hepatitis B surface antigen （HBsAg）-positive and HBsAg-negative/hepatitis B core antibody-positive patients was 13% （3/24） and 4% （3/69）, respectively. Conclusions： R-chemo is effective and safe in real-world clinical practice as first-line treatment for DLBCL and FL in China, and that HBV reactivation during R-chemo is manageable wi
文摘研究RoboCup比赛未知环境中足球机器人的路径规划问题。提出一种多优化设计快速扩展随机树(rapidly exploring random tree,RRT)的足球机器人路径规划算法,解决了足球机器人在路径规划中存在的速度慢、效果差的问题。依据基本RRT算法原理,针对其随机性强、收敛速度慢以及路径平滑效果差的缺点,提出了随机采样点处增加引力分量、多步扩展逼近目标点以及冗余节点的剪裁与路径平滑等优化方式。在不同障碍物地图中的仿真实验表明,优化的规划路径长度比基本快速扩展随机树算法所得路径缩短约20%~30%,采样点数量减少45%~65%。最终将优化算法移植到SimRobot仿真平台,结果表明多优化设计RRT算法在未知环境中具备良好的实时性和鲁棒性,能够满足机器人在赛场上的路径规划要求。
