I am trying to calculate SSIM for 2 numpy arrays using structural_similarity from skimage.metrics:
ydata.shape = (3, 230)
ydata3.shape = (3, 230)
When I try to calculate SSIM:
mse_value = mean_squared_error(ydata, ydata3)
ssim_value = ssim(ydata, ydata3, data_range=ydata3.max() - ydata.min())**
I have the error:
"win_size exceeds image extent. If the input is a multichannel "
ValueError: win_size exceeds image extent. If the input is a multichannel (color) image, set multichannel=True.**
I tried to set multichannel = True but it does not work.
ssm: ssim_value = ssim(ydata, ydata3, multichannel= True, data_range=ydata3.max() - ydata.min())
I also tried to increase the dimension of the array with:
ydata = ydata.reshape(ydata.shape[0], ydata.shape[1], 1)
The new shape of the arrays is ((3, 230, 1), (3, 230, 1)). However it keeps showing the same error
I tried also flatten but it changes a lot my results.
Any suggestions or help?