SRFeat: Single Image Super-Resolution with Feature Discrimination
發表自ECCV2018. 主要工作： 1.使用了兩個判別器，分別作用於圖像和特徵圖。 2. 提出了具有遠程跳躍連接的生成器。 GAN: 生成器損失： 判別器損失 : 其中表示判別器的輸出，也就是圖像是從真實的圖像分佈中進行採樣得到的概率。 我們知道log函數在底數大於1時是增函數，所以對於生成器，必定是希望生成的圖像被判別器判別爲真實的圖像的概率越大越好，所以取負數以後損失變小了，同理可得判別器
相關文章
- 1. 文獻閱讀:SRFeat: Single Image Super-Resolution with Feature Discrimination
- 2. Natural and Realistic Single Image Super-Resolution with Explicit Natural Manifold Discrimination
- 3. Multi-Level Feature Fusion Mechanism for Single Image Super-Resolution
- 4. [解讀] Unsupervised Feature Learning via Non-Parametric Instance Discrimination
- 5. AIPNet: Image-to-Image Single Image Dehazing with Atmospheric Illumination Prior
- 6. Hyperspectral Image Classification With Deep Feature Fusion Network
- 7. 《Real Image Denoising with Feature Attention》閱讀筆記
- 8. Unsupervised Feature Learning via Non-Parametric Instance Discrimination
- 9. Unsupervised Feature Learning via Non-parametric Instance Discrimination
- 10. Guided Image-to-Image Translation with Bi-Directional Feature Transformation（ICCV19）
- 更多相關文章...
- • RSS
元素 - RSS 教程 - • ASP.NET Image 控件 - ASP.NET 教程
- • RxJava操作符（十）自定義操作符
- • Docker 清理命令
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。
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. 文獻閱讀:SRFeat: Single Image Super-Resolution with Feature Discrimination
- 2. Natural and Realistic Single Image Super-Resolution with Explicit Natural Manifold Discrimination
- 3. Multi-Level Feature Fusion Mechanism for Single Image Super-Resolution
- 4. [解讀] Unsupervised Feature Learning via Non-Parametric Instance Discrimination
- 5. AIPNet: Image-to-Image Single Image Dehazing with Atmospheric Illumination Prior
- 6. Hyperspectral Image Classification With Deep Feature Fusion Network
- 7. 《Real Image Denoising with Feature Attention》閱讀筆記
- 8. Unsupervised Feature Learning via Non-Parametric Instance Discrimination
- 9. Unsupervised Feature Learning via Non-parametric Instance Discrimination
- 10. Guided Image-to-Image Translation with Bi-Directional Feature Transformation（ICCV19）