DL去燥之 DnCNNs:Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
2017-06-09 00:02
686 查看
1、网络结构:
DnCNNs: feed-forward denoising convolutional neural networks2、学习策略
Residual Learning:通过残差学习策略,在网络的隐层隐式地移除干净图片(clean image)。即输入为带噪的观测图片(noisy image),输出为移除了干净图片的残差(噪声)图片(residual image)。这样做的动机是,用残差学习刻画恒等映射或近似的恒等映射,效果要比直接学习干净图片要相对好。Batch Normalization:进一步提高模型的稳定性和去燥效果。
3、适用范围
Gaussian Denoising:DnCNN model is able to handle Gaussian denoising with unknown noise level (i.e., blind Gaussian denoising);Image Super-resolution;
JPEG Deblocking.
相关文章推荐
- 《Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising》学习笔记
- 看讲论文:Learning Deep CNN Denoiser Prior for Image Restoration
- CNN初探(四)------Deep Residual Learning for Image Recognition
- 论文阅读-Learning Deep CNN Denoiser Prior for Image Restoration
- Deep Residual Learning for Image Recognition笔记
- Caffe build for denoising of deep learning
- Deep Residual Learning for Image Recognition
- 翻译:Deep Residual Learning for Image Recognition
- 基于deep learning的快速图像检索(Deep Learning of Binary Hash Codes for Fast Image Retrieval)
- 图像识别的深度残差学习Deep Residual Learning for Image Recognition
- Deep Residual Learning for Image Recognition 笔记
- Deep Learning of Binary Hash Codes for Fast Image Retrieval
- Deep Residual Learning for Image Recognition
- Deep Residual Learning for Image Recognition
- deep learning Deep Belief networks for image denoising
- [Editing] Deep Residual Learning for Image Recognition
- Deep Residual Learning for Image Recognition
- 基于deep learning的快速图像检索(Deep Learning of Binary Hash Codes for Fast Image Retrieval)
- Deep Residual Learning for Image Recognition(译)
- Deep Residual Learning for Image Recognition