論文筆記Globally and Locally Consistent Image Completion

論文筆記Globally and Locally Consistent Image Completion 圖1 Patch-based方法的示例圖 Patch-based方法的不足:(1)Depend on low-level features;(2)Unable to generate novel objects;2)基於Context Encoder的圖像補全,該方法基於深度學習生成相似的紋理
相關文章
相關標籤/搜索