The visual automatic detection method based on artificial intelligence has attracted more and more attention. In order to improve the performance of weld nondestructive defect detection,we propose DRepDet(Dilated RepP...The visual automatic detection method based on artificial intelligence has attracted more and more attention. In order to improve the performance of weld nondestructive defect detection,we propose DRepDet(Dilated RepPoints Detector). First, we analyze the weld defect dataset in detail and summarize the distribution characteristics of weld defect data, that is, the defect scale is very different and the aspect ratio distribution range is large. Second, according to the distribution characteristics of defect data, we design DResBlock module, and introduce dilated convolution with different dilated rates in the process of feature extraction to expand the receptive field and improve the detection performance of large-scale defects. Based on DResBlock and anchor-free detection framework RepPoints, we design DRepDet. Extensive experiments show that our proposed detector can detect 7 types of defects. When using combined dilated rate convolution network in detection, the AP50 and Recall50 of big defects are improved by 3.1% and 3.3% respectively, while the performance of small defects is not affected, almost the same or slightly improved. The final performance of the whole network is improved a large margin,with 6% AP50 and 4.2% Recall50 compared with Cascade RCNN and 1.4% AP50 and 2.9% Recall50 compared with RepPoints.展开更多
Community smells are sub-optimal developer community structures that hinder productivity.Prior studies performed smell prediction and provided refactoring guidelines from a top-down aspect to help community shepherds....Community smells are sub-optimal developer community structures that hinder productivity.Prior studies performed smell prediction and provided refactoring guidelines from a top-down aspect to help community shepherds.Simultaneously,refactoring smells also requires bottom-up effort from every developer.However,supportive measures and guidelines for them are not available at a fine-grained level.Since recent work revealed developers'personalities and working states could influence community smells'emergence and variation,we build prediction models with experience,sentiment,and development process features of developers considering three smells including Organizational Silo,Lone Wolf,and Bottleneck,as well as two related classes including smelly developer and smelly quitter.We predict the five classes in the individual granularity,and we also generate forecasts for the number of smelly developers in the community granularity.The proposed models achieve F-measures ranging from 0.73 to 0.92 in individual-wide within-project,time-wise,and cross-project prediction,and mean R2 performance of 0.68 in community-wide Smelly Developer prediction.We also exploit SHAP(SHapley Additive exPlanations)to assess feature importance to explain our predictors.In conclusion,we suggest developers with heavy workload should foster more frequent communication in a straightforward and polite way to build healthier communities,and we recommend community shepherds to use the forecasting model for refactoring planning.展开更多
文摘The visual automatic detection method based on artificial intelligence has attracted more and more attention. In order to improve the performance of weld nondestructive defect detection,we propose DRepDet(Dilated RepPoints Detector). First, we analyze the weld defect dataset in detail and summarize the distribution characteristics of weld defect data, that is, the defect scale is very different and the aspect ratio distribution range is large. Second, according to the distribution characteristics of defect data, we design DResBlock module, and introduce dilated convolution with different dilated rates in the process of feature extraction to expand the receptive field and improve the detection performance of large-scale defects. Based on DResBlock and anchor-free detection framework RepPoints, we design DRepDet. Extensive experiments show that our proposed detector can detect 7 types of defects. When using combined dilated rate convolution network in detection, the AP50 and Recall50 of big defects are improved by 3.1% and 3.3% respectively, while the performance of small defects is not affected, almost the same or slightly improved. The final performance of the whole network is improved a large margin,with 6% AP50 and 4.2% Recall50 compared with Cascade RCNN and 1.4% AP50 and 2.9% Recall50 compared with RepPoints.
基金supported by the National Natural Science Foundation of China under Grant No.61772200the Natural Science Foundation of Shanghai under Grant No.21ZR1416300.
文摘Community smells are sub-optimal developer community structures that hinder productivity.Prior studies performed smell prediction and provided refactoring guidelines from a top-down aspect to help community shepherds.Simultaneously,refactoring smells also requires bottom-up effort from every developer.However,supportive measures and guidelines for them are not available at a fine-grained level.Since recent work revealed developers'personalities and working states could influence community smells'emergence and variation,we build prediction models with experience,sentiment,and development process features of developers considering three smells including Organizational Silo,Lone Wolf,and Bottleneck,as well as two related classes including smelly developer and smelly quitter.We predict the five classes in the individual granularity,and we also generate forecasts for the number of smelly developers in the community granularity.The proposed models achieve F-measures ranging from 0.73 to 0.92 in individual-wide within-project,time-wise,and cross-project prediction,and mean R2 performance of 0.68 in community-wide Smelly Developer prediction.We also exploit SHAP(SHapley Additive exPlanations)to assess feature importance to explain our predictors.In conclusion,we suggest developers with heavy workload should foster more frequent communication in a straightforward and polite way to build healthier communities,and we recommend community shepherds to use the forecasting model for refactoring planning.