pyiqa.archs.gmsd_arch

GMSD Metric

@article{xue2013gmsd,

title={Gradient magnitude similarity deviation: A highly efficient perceptual image quality index}, author={Xue, Wufeng and Zhang, Lei and Mou, Xuanqin and Bovik, Alan C}, journal={IEEE transactions on image processing}, volume={23}, number={2}, pages={684–695}, year={2013}, publisher={IEEE}

}

Created by: https://github.com/dingkeyan93/IQA-optimization/blob/master/IQA_pytorch/GMSD.py

Modified by: Jiadi Mo (https://github.com/JiadiMo)

Refer to:

Matlab code from https://www4.comp.polyu.edu.hk/~cslzhang/IQA/GMSD/GMSD.m;

Module Contents

pyiqa.archs.gmsd_arch.gmsd(x: torch.Tensor, y: torch.Tensor, T: int = 170, channels: int = 3, test_y_channel: bool = True) → torch.Tensor

GMSD metric. :param - x: A distortion tensor. Shape \((N, C, H, W)\). :param - y: A reference tensor. Shape \((N, C, H, W)\). :param - T: A positive constant that supplies numerical stability. :param - channels: Number of channels. :param - test_y_channel: bool, whether to use y channel on ycbcr.

class pyiqa.archs.gmsd_arch.GMSD(channels: int = 3, test_y_channel: bool = True)

Bases: torch.nn.Module

Gradient Magnitude Similarity Deviation Metric. :param - channels: Number of channels. :param - test_y_channel: bool, whether to use y channel on ycbcr.

Reference:

Xue, Wufeng, Lei Zhang, Xuanqin Mou, and Alan C. Bovik. “Gradient magnitude similarity deviation: A highly efficient perceptual image quality index.” IEEE Transactions on Image Processing 23, no. 2 (2013): 684-695.

forward(x: torch.Tensor, y: torch.Tensor) → torch.Tensor

Args: x: A distortion tensor. Shape \((N, C, H, W)\). y: A reference tensor. Shape \((N, C, H, W)\). Order of input is important.