Comments (1)
非常感谢您的提问。首先就论文写作不足致歉。关于每个疑问的回复如下:
-
这个问题还是因为你对Meta-Bicu不理解。假设你将一个图片放大四倍,再用3 * 3的卷积去卷积的时候,它只用到了原图上周围2 * 2的点,插值1.5倍时,本质用到原图3*3上的点。FOV的定义我就不过多介绍了。
-
所有baselines都是在多scale factor上训练的单模型。
-
一个应该是歧义导致的,文章中的these two methods是指EDSRx1和RDNx1,不过导致歧义确实也需要修改。第二个,论文里没错吧。我就是这么写的啊。您在仔细看看。
-
不在于多尺度(scale factor?)吧。我们的Meta-SR是可以在单尺度(scale factor)上训练的;还是在于基于卷积放大的方法要优于插值放大,再利用meta-learning。这两者一起才是这个方法优于其他方法的原因。您可以比较Meta-Conv 和 Meta-SR,这里说明了插值的放大的方法比不过基于卷积放大。
你提的实验文章都给出了吧?
在所有比较的baseline中,它们都是多尺度训练的,既然如此为什么考虑这个?这是一个不变量。
注:文章的目的是要设计一个算法,实现单模型任意放大倍数。 -
您这种做法(应该称为Meta-nearest)应该还不如我的Meta-Bicu,本质上改变的只有插值放大的模式,采用bicubic还是nearest。至于如果您是考虑浪费了很多额外的运算的话, 主要这是我能想到最合理的矩阵实现方式。如果您有更好的,您可以提供伪代码。我来实现。
from meta-sr-pytorch.
Related Issues (20)
- Have you debugged it yet?
- Meta-Upscale Module
- meta-upscale
- meta-upscale的输入
- 请问输入矩阵为什么需要mask
- Meta-upscale的实现 HOT 3
- RuntimeError: cuda runtime error (2) HOT 5
- Trying to train Meta-RCAN but failed HOT 2
- Testing directories HOT 3
- rewrite dataloader for more recnt Pytorch
- meta-learning for weight prediction
- dataloader error, help plz~
- Higher PSNR when i use pretrained model?
- 请问怎样运行 geberate_LR_metasr_X1_X4.m 文件?
- Pretrained models
- 如何将MetaUpSampler 改成适用于3d图像的上采样?
- 请问,想改成 针对3d数据,该怎么改? 比如(batch,C, h, w, d),超分到(batch, C, H, W, D)。 HOT 2
- 你好,能帮忙指点下吗? 改成3d 后 pos_mat_small 维度不是Scale x Scale x Scale x 3的维度? h_offset这需要改吗? HOT 3
- 你好,cols = nn.functional.unfold(up_x.permute(0, 2, 3, 1), self.kernel_size, padding=1),该咋改呀? HOT 1
- Pre-training model selection for testing
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from meta-sr-pytorch.