論文閱讀——Fully Convolutional Architectures for Multi-Class Segmentation in Chest Radiographs

論文閱讀之InvertedNet Fully Convolutional Architectures for Multi-Class Segmentation in Chest Radiographs from TMI2018 摘要： 本文提出的網絡結合了delayed subsampling、指數線性單元（exponential linear unit)、以及大量高分辨率的淺層特徵用於胸片分割（

>>阅读原文<<