-
I can't find the test dataset in the authors' website.Can you provide the test dataset and the code of PSNR and SSIM?Thank you very much!
-
**Is your feature request related to a problem? Please describe.**
Proposed in 2016 this method shows good performance in general image domain.
**Describe the solution you'd like**
Implement it i…
-
论文LOL_v2_real中的psnr和ssim分别为22.00和0.849，预训练模型的测试结果为22.45和0.844
-
A followup to #48, we should follow @pahammad's advice and do what visualmetrics does, using SSIM as the progress once the first paint has occurred and do not normalize.
The normalization causes pr…
-
Part 2 of the big performance improvements for jest-image-snapshot's implementation of SSIM.
In the latest release of SSIM (4.2.0) we made some major improvements to the implementation that increas…
-
I am using the default target of `0.9999` and the `ms-ssim` method, but I've noticed the output quality is always what the `min` setting is set to, as follows:
```
# ./jpeg-recompress --min 50 --met…
-
感谢作者的贡献！ECBSR模型在DIV2K的bicubic的数据上PSNR能确实到达很高。为了更好的落地作者的算法，我们尝试用同样的模型训练一批仿真的DIV2K多帧Y数据（退化方案不仅是Bicubic，而是复杂多样的，并通过亚像素位移融入多帧信息），得出来的PSNR结果却降到26左右且很难上升到30以上，Loss也很难像bicubic一样降到3左右：
##===========Epoch: 8=…
-
Currently SSIM is supported: https://kornia.readthedocs.io/en/latest/losses.html#kornia.losses.SSIM
Multiscale SSIM would also be nice.
example: https://github.com/jorge-pessoa/pytorch-msssim
-
It seems Pivia Center dataset preprocessing program is not exists. Can you provide it ?
-
I have done some tests with the last AVX version and I have found that in most tests, it has different VMAF metrics than plain version:
![image](https://github.com/Patman86/x265-Mod-by-Patman/asset…