【論文閱讀筆記】Multi-scale context aggregation by dilated convolutions

論文地址:https://arxiv.org/abs/1511.07122 代碼地址:https://github.com/ndrplz/dilation-tensorflow https://github.com/fyu/dilation 作者用caffe寫的 introduction motivation 動機 CNN的池化操作影響語義分割的精度 This prompts new questi
相關文章
相關標籤/搜索