Paper | U-Net: Convolutional Networks for Biomedical Image Segmentation

目錄 故事背景 U-Net 具體結構 損失 數據擴充 發表在2015 MICCAI。原本是一篇醫學圖像分割的論文，但由於U-Net傑出的網絡設計，得到了8k+的引用。 摘要 There is large consent that successful training of deep networks requires many thousand annotated training sample

>>阅读原文<<