PSNR error with operands, different shapes

Question

hi guys

am using a code to apply a Wiener filter on a noisy image, the code source

and it worked and everything seemed to be okay until I went to measure the PSNR value, and did not work, I used this method to measure PSNR :

 def PSNR(original, compressed):
    mse = np.mean((original - compressed) ** 2)
    if(mse == 0): 
        return d
    max_pixel = 255.0
    psnr = 20 * log10(max_pixel / sqrt(mse))
    return psnr


original = img
compressed = filtered_img
value = PSNR(original, compressed)
print("noisy = ",value)

The error message is :

ValueError                                Traceback (most recent call last)
<ipython-input-18-f21fe64cbb4c> in <module>()
     73 original = img
     74 compressed = filtered_img
---> 75 value = PSNR(original, compressed)
     76 print("noisy = ",value)

<ipython-input-6-7d8e6558cdd2> in PSNR(original, compressed)
      2 
      3 def PSNR(original, compressed):
----> 4     mse = np.mean((original - compressed) ** 2)
      5     if(mse == 0):
      6         return d

ValueError: operands could not be broadcast together with shapes (512,512) (498,498) 

I want to use this code to denoise an image with speckle or Gaussian noise, but I need PSNR calculation to identify or measure the denoising process.

so pls help me fix this problem, thanks in advance