Retinal vessel segmentation is a significant problem in the analysis of fundus images.A novel deep learning structure called the Gaussian net(GNET)model combined with a saliency model is proposed for retinal vessel se...Retinal vessel segmentation is a significant problem in the analysis of fundus images.A novel deep learning structure called the Gaussian net(GNET)model combined with a saliency model is proposed for retinal vessel segmentation.A saliency image is used as the input of the GNET model replacing the original image.The GNET model adopts a bilaterally symmetrical structure.In the left structure,the first layer is upsampling and the other layers are max-pooling.In the right structure,the final layer is max-pooling and the other layers are upsampling.The proposed approach is evaluated using the DRIVE database.Experimental results indicate that the GNET model can obtain more precise features and subtle details than the UNET models.The proposed algorithm performs well in extracting vessel networks,and is more accurate than other deep learning methods.Retinal vessel segmentation can help extract vessel change characteristics and provide a basis for screening the cerebrovascular diseases.展开更多
Background Although it is generally acknowledged that patients with ruptured abdominal aortic aneurysm （rAAA） obtain the greatest benefit from endovascular repair （EVAR）, convincing evidence on the medium-long ter...Background Although it is generally acknowledged that patients with ruptured abdominal aortic aneurysm （rAAA） obtain the greatest benefit from endovascular repair （EVAR）, convincing evidence on the medium-long term effect is lacking. The aim of this study was to compare and summarize published results of rAAA that underwent EVAR with open surgical repair （OSR）. Methods A search of publicly published literature was performed. Based on an inclusion and exclusion criteria, a systematic meta-analysis was undertaken to compare patient characteristics, complications, short term mortality and medium-long term outcomes. A random-effects model was used to pool the data and calculate pooled odds ratios and weighted mean differences. A quantitative method was used to analyze the differences between these two methods. Results A search of the published literature showed that fourteen English language papers comprising totally 1213 patients with rAAA （435 EVAR and 778 OSR） would be suitable for this study. Furthermore, 13 Chinese studies were included, including 267 patients with rAAA totally, among which 238 patients received operation. The endovascular method was associated with more respiratory diseases before treatment （OR=1.81, P=0.01）, while there are more patients with hemodynamic instability before treatment in OSR group （OR=1.53, P=0.031）. Mean blood transfusion was 1328 ml for EVAR and 2809 ml for OSR （weighted mean difference （WMD） 1500 ml, P=0.014）. The endovascular method was associated with a shorter stay in intensive care （WMD 2.34 days, P 〈0.001） and a shorter total post- operative stay （WMD 6.27 days, P 〈0.001）. The pooled post-operative complication rate of respiratory system and visceral ischemia seldom occurred in the EVAR group （OR=0.48, P 〈0.001 and OR=0.28, P=0.043, respectively）. The pooled 30-day mortality was 25.7% for EVAR and 39.6% for OSR, and the odds ratio was 0.53 （95% confidence interval （CI） 0.41-0.70, P 〈0.001）. There was not, however, any significant reduction in the medium-long all-cause mortality rate （HR=1.13, P=0.381） and re-intervention rate （OR= 2.19, ,~=-0.243） following EVAR. In EVAR group, nevertheless, incidence of type I endoleak was significantly lower than type II endoleak （OR=0.33, P=0.039） at late follow-up period. Conclusions On the basis of this systematic review, rAAA EVAR results in less blood use for transfusion, shorter operation time, shorter intensive care unit and hospital stays, and lower 30-day mortality. However, in the medium-long term, it is not associated with a reduction in all-cause mortality.展开更多
基金Project supported by the Natural Science Foundation of Fujian Province,China(No.2016J0129)the Educational Commission of Fujian Province of China(No.JAT170180)
文摘Retinal vessel segmentation is a significant problem in the analysis of fundus images.A novel deep learning structure called the Gaussian net(GNET)model combined with a saliency model is proposed for retinal vessel segmentation.A saliency image is used as the input of the GNET model replacing the original image.The GNET model adopts a bilaterally symmetrical structure.In the left structure,the first layer is upsampling and the other layers are max-pooling.In the right structure,the final layer is max-pooling and the other layers are upsampling.The proposed approach is evaluated using the DRIVE database.Experimental results indicate that the GNET model can obtain more precise features and subtle details than the UNET models.The proposed algorithm performs well in extracting vessel networks,and is more accurate than other deep learning methods.Retinal vessel segmentation can help extract vessel change characteristics and provide a basis for screening the cerebrovascular diseases.
基金This work was supported by Science Foundation of China grants from the National Natural （No. 304717076）, the Department of Education of Liaoning Province （Key Laboratory Project No. LS2010172）, and Ministry of Education of China （Key Research Project of Science and Technology No. 208028）.
文摘Background Although it is generally acknowledged that patients with ruptured abdominal aortic aneurysm （rAAA） obtain the greatest benefit from endovascular repair （EVAR）, convincing evidence on the medium-long term effect is lacking. The aim of this study was to compare and summarize published results of rAAA that underwent EVAR with open surgical repair （OSR）. Methods A search of publicly published literature was performed. Based on an inclusion and exclusion criteria, a systematic meta-analysis was undertaken to compare patient characteristics, complications, short term mortality and medium-long term outcomes. A random-effects model was used to pool the data and calculate pooled odds ratios and weighted mean differences. A quantitative method was used to analyze the differences between these two methods. Results A search of the published literature showed that fourteen English language papers comprising totally 1213 patients with rAAA （435 EVAR and 778 OSR） would be suitable for this study. Furthermore, 13 Chinese studies were included, including 267 patients with rAAA totally, among which 238 patients received operation. The endovascular method was associated with more respiratory diseases before treatment （OR=1.81, P=0.01）, while there are more patients with hemodynamic instability before treatment in OSR group （OR=1.53, P=0.031）. Mean blood transfusion was 1328 ml for EVAR and 2809 ml for OSR （weighted mean difference （WMD） 1500 ml, P=0.014）. The endovascular method was associated with a shorter stay in intensive care （WMD 2.34 days, P 〈0.001） and a shorter total post- operative stay （WMD 6.27 days, P 〈0.001）. The pooled post-operative complication rate of respiratory system and visceral ischemia seldom occurred in the EVAR group （OR=0.48, P 〈0.001 and OR=0.28, P=0.043, respectively）. The pooled 30-day mortality was 25.7% for EVAR and 39.6% for OSR, and the odds ratio was 0.53 （95% confidence interval （CI） 0.41-0.70, P 〈0.001）. There was not, however, any significant reduction in the medium-long all-cause mortality rate （HR=1.13, P=0.381） and re-intervention rate （OR= 2.19, ,~=-0.243） following EVAR. In EVAR group, nevertheless, incidence of type I endoleak was significantly lower than type II endoleak （OR=0.33, P=0.039） at late follow-up period. Conclusions On the basis of this systematic review, rAAA EVAR results in less blood use for transfusion, shorter operation time, shorter intensive care unit and hospital stays, and lower 30-day mortality. However, in the medium-long term, it is not associated with a reduction in all-cause mortality.
