iizhuzhu的博客

目标检测和分割中Backbone技术总结VGG (2014)ResNet系列ResNet (2015)ResNeXtResNeStInception系列Inception v1Inception v2Inception v3Inception V4总结SENet系列SENetSKNet (Selective Kernel Networks)U-NetFeature Pyramid Network (FPN)introductionFPNFaster R-CNN with FPNDCNDCN v1Deform

abstract: |Compared with MS-COCO, the dataset for the competition has a largerproportion of large objects which area is greater than 96$\times$96pixels. As getting fine boundaries is vitally important for large objectsegmentation, Mask R-CNN with Poin.

Content1 Introduction2 OverView3 Input Data3.1 Data Augmentation3.2 Sliding Sample4 Model4.1 Model Zoo4.2 Loss Function4.3 Learn Rate Scheduler4.3.1 torch.optim.lr_scheduler.ReduceLROnPlateau4.3.2 tor...