[Paper Note] Meta-SR A Magnification-Arbitrary Network for Super-Resolution
Summary 這篇文章的創新點在於提出了Meta-Upscale Module這個模塊，實現了任意尺寸的放大（當然這篇論文做的主要是x1-x4的倍數），核心思想是HR和LR圖片之間實際上是一個縮放關係，有一個縮放因子r。LR的像素通過一組濾波器權重可以映射到SR上，所以找到SR圖片和LR圖片的像素映射，訓練出濾波器權重。 (1) I S R ( i , j ) = Φ ( F L R ( i ′
相關文章
- 1. [Paper Reading Note]Path Aggregation Network for Instance Segmentation
- 2. Reading Note: DetNet: A Backbone network for Object Detection
- 3. Paper reading：PointGrid： A Deep Network for 3D Shape Understanding
- 4. SSD paper note
- 5. [Paper][Link note]
- 6. [Paper Reading Note]Aggregated Residual Transformations for Deep Neural Networks
- 7. Paper Reading: Attention-Aware Compositional Network for Person Re-identification
- 8. 【Paper Note】xDeepFM詳解
- 9. A brief scanning of paper 「Spatial Transformer Network」
- 10. Reading Note: ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression
- 更多相關文章...
- • Scala for循環 - Scala教程
- • Lua for 循環 - Lua 教程
- • Flink 數據傳輸及反壓詳解
- • Kotlin學習（一）基本語法
相關標籤/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
最新文章
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. [Paper Reading Note]Path Aggregation Network for Instance Segmentation
- 2. Reading Note: DetNet: A Backbone network for Object Detection
- 3. Paper reading：PointGrid： A Deep Network for 3D Shape Understanding
- 4. SSD paper note
- 5. [Paper][Link note]
- 6. [Paper Reading Note]Aggregated Residual Transformations for Deep Neural Networks
- 7. Paper Reading: Attention-Aware Compositional Network for Person Re-identification
- 8. 【Paper Note】xDeepFM詳解
- 9. A brief scanning of paper 「Spatial Transformer Network」
- 10. Reading Note: ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression