Hi, first of all thank you for sharing such a great tool.

i've tried to install by using pip but i got following error. Could you help?

PS E:\MyAllProjects\Python\BoundingBox> pip install image-similarity-measures
Collecting image-similarity-measures
Using cached image_similarity_measures-0.3.5-py3-none-any.whl (9.1 kB)
Requirement already satisfied: numpy in e:\myallprojects\python\boundingbox\dataprep\lib\site-packages (from image-similarity-measures) (1.22.3)
Collecting rasterio
Using cached rasterio-1.2.10.tar.gz (2.3 MB)
Installing build dependencies ... done
Getting requirements to build wheel ... error
error: subprocess-exited-with-error

× Getting requirements to build wheel did not run successfully.
│ exit code: 1
╰─> [2 lines of output]
INFO:root:Building on Windows requires extra options to setup.py to locate needed GDAL files. More information is available in the README.
ERROR: A GDAL API version must be specified. Provide a path to gdal-config using a GDAL_CONFIG environment variable or use a GDAL_VERSION environment variable.
[end of output]

note: This error originates from a subprocess, and is likely not a problem with pip.
error: subprocess-exited-with-error

× Getting requirements to build wheel did not run successfully.
│ exit code: 1
╰─> See above for output.

note: This error originates from a subprocess, and is likely not a problem with pip.

I checked the RMSE computing process and it returns a mean, but it should be np.sqrt at last.

1.https://docs.opencv.org/4.x/dd/d3d/tutorial_gpu_basics_similarity.html
`

double getPSNR(const Mat& I1, const Mat& I2)
{
Mat s1;
absdiff(I1, I2, s1); // |I1 - I2|
s1.convertTo(s1, CV_32F); // cannot make a square on 8 bits
s1 = s1.mul(s1); // |I1 - I2|^2
Scalar s = sum(s1); // sum elements per channel
double sse = s.val[0] + s.val[1] + s.val[2]; // sum channels
if( sse <= 1e-10) // for small values return zero
return 0;
else
{
double mse =sse /(double)(I1.channels() * I1.total());
double psnr = 10.0log10((255255)/mse);
return psnr;
}
}
2. image-similarity-measuresdef psnr(org_img: np.ndarray, pred_img: np.ndarray, max_p=4095) -> float:
"""
Peek Signal to Noise Ratio, implemented as mean squared error converted to dB.

It can be calculated as
PSNR = 20 * log10(MAXp) - 10 * log10(MSE)

When using 12-bit imagery MaxP is 4095, for 8-bit imagery 255. For floating point imagery using values between
0 and 1 (e.g. unscaled reflectance) the first logarithmic term can be dropped as it becomes 0
"""
_assert_image_shapes_equal(org_img, pred_img, "PSNR")

org_img = org_img.astype(np.float32)

mse_bands = []
for i in range(org_img.shape[2]):
    mse_bands.append(np.mean(np.square(org_img[:, :, i] - pred_img[:, :, i])))

return 20 * np.log10(max_p) - 10. * np.log10(np.mean(mse_bands))`

the same equation (10.0*log10((255 * 255)/mse) = 20 * log10(MAXp) - 10 * log10(MSE))， but get different result . i don't know why ?

the value is set to 12 bits (4095)
I can't understand it. Where should I set this in the code?

When I run the code, it fails. 'image_similarity_measures' is a folder name，what should I do?

Hi there!

I had a quick question. The skimage.metrics also has a structural similarity index measure. If I apply yours and the one of skimage to the same image, I get different results. Do you might know why this is the case?

Kind regards!

I've noticed that the SSIM and PSNR values are much higher when using your tool than when trying ffmpeg method such as:
ffmpeg -i first.png -i second.png -lavfi ssim -f null -
Why is that?
Thanks for the tool anyway ^_^.

Can We Use FSIM with a normalized image between [0 1] range because I have normalized images

How to use this code for different image dimension

Input 256x256
output 1024x1024 after super resolution

Hi,

I am wondering how to avoid it saves the result into a file.
I have try different values but it keeps saving files

--write_to_file='FALSE'
--write_to_file=FALSE
--write_to_file=False

Could you indicate me what is the right parameter value?

Thanks in advance, JL

It looks like project is needs simple rebuild to work with Python 3.10 and Python 3.11

Python3.11:

(venv) [nikow@thinira tmp]$ python --version
Python 3.11.1
(venv) [nikow@thinira tmp]$ python -m pip install --prefer-binary image-similarity-measures
ERROR: Ignored the following versions that require a different python version: 0.0.1 Requires-Python >=3.6, <3.8; 0.1.1 Requires-Python >=3.6, <3.8; 0.1.2 Requires-Python >=3.6, <3.8; 0.2.2 Requires-Python >=3.6, <3.8; 0.3.0 Requires-Python >=3.6, <3.8; 0.3.1 Requires-Python >=3.6, <3.8; 0.3.3 Requires-Python >=3.6, <3.8; 0.3.4 Requires-Python >=3.6, <3.9; 0.3.5 Requires-Python >=3.6, <3.10
ERROR: Could not find a version that satisfies the requirement image-similarity-measures (from versions: none)
ERROR: No matching distribution found for image-similarity-measures
(venv) [nikow@thinira tmp]$

Python3.10:

(venv) [nikow@thinira tmp]$ python --version
Python 3.10.9
(venv) [nikow@thinira tmp]$ python -m pip install image-similarity-measures
ERROR: Ignored the following versions that require a different python version: 0.0.1 Requires-Python >=3.6, <3.8; 0.1.1 Requires-Python >=3.6, <3.8; 0.1.2 Requires-Python >=3.6, <3.8; 0.2.2 Requires-Python >=3.6, <3.8; 0.3.0 Requires-Python >=3.6, <3.8; 0.3.1 Requires-Python >=3.6, <3.8; 0.3.3 Requires-Python >=3.6, <3.8; 0.3.4 Requires-Python >=3.6, <3.9; 0.3.5 Requires-Python >=3.6, <3.10
ERROR: Could not find a version that satisfies the requirement image-similarity-measures (from versions: none)
ERROR: No matching distribution found for image-similarity-measures
(venv) [nikow@thinira tmp]$ 

The following error shows up and returns ISSM=0

site-packages/image_similarity_measures/quality_metrics.py:146:
RuntimeWarning: divide by zero encountered in log2
return -np.sum(np.nan_to_num(H * np.log2(H)))

Hi! Im developing my own image format hence i need this package for lossy compression quality evaluation. Can you please update this package for python 3.10 usage. Thanks.

p.s.
All dependencies support python 3.10 so now i just copied quality_metrics.py into my project.

I computed the SAM value for 2 RGB images, but then I realized it seems only can be used to compute spectrum graph...But I still got the result of SAM Value during [89,90].
Is that a mistake? Hope for answer.

Hi,

I'm trying to use this package but pip install image-similarity-measures doesn't seem to install a binary (I'm using pyenv).

~ pip install image-similarity-measures
...
~ image-similarity-measures
zsh: command not found: image-similarity-measures

I'm sure I'm missing something basic (I'm a python noob). Am I doing something wrong?

~ pip list
Package                   Version
------------------------- ---------
image-similarity-measures 0.3.6
imageio                   2.31.1
lazy_loader               0.3
networkx                  3.1
numpy                     1.25.1
opencv-python             4.8.0.74
packaging                 23.1
phasepack                 1.5
Pillow                    10.0.0
pip                       23.1.2
PyWavelets                1.4.1
scikit-image              0.20.0
scipy                     1.11.1
setuptools                65.5.0
tifffile                  2023.7.10

Thanks!

Thank you for providing this tool.
It works from the command line in Ubuntu 20.04 which is great.

I am trying to make image_similarity_measures work from within a python module and have tried various approaches, but I feel the Usage in python instructions could benefit with a more detailed example.

These were my attempts below:

My imports
import image_similarity_measures
from image_similarity_measures.quality_metrics import fsim, issm, psnr, rmse, sam, sre, ssim, uiq
from image_similarity_measures.evaluate import evaluation # I tried adding this import after other attempts

My failed attempts
res = image_similarity_measures("--org_img_path=tmp1.png --pred_img_path=tmp2.png --metric=all")
res = image_similarity_measures.evaluate("tmp1.png", "tmp2.png", "all")
res = image_similarity_measures.evaluate.evaluation("tmp1.png", "tmp2.png", "all")
res = rmse("images/tmp1.png", "images/tmp2.png")

Describe the bug
A clear and concise description of what the bug is.

To Reproduce
Steps to reproduce the behavior:

1. execute pip install image-similarity-measures on Windows 10.
2. See error

Collecting image-similarity-measures
  Downloading image_similarity_measures-0.3.5-py3-none-any.whl (9.1 kB)
Requirement already satisfied: scikit-image in e:\software\anaconda3\lib\site-packages (from image-similarity-measures) (0.17.2)
Collecting rasterio
  Downloading rasterio-1.2.3.tar.gz (2.2 MB)
     |████████████████████████████████| 2.2 MB 504 kB/s
  Installing build dependencies ... done
  Getting requirements to build wheel ... error
  ERROR: Command errored out with exit status 1:
   command: 'E:\SoftWare\Anaconda3\python.exe' 'E:\SoftWare\Anaconda3\lib\site-packages\pip\_vendor\pep517\_in_process.py' get_requires_for_build_wheel 'C:\Users\ZHANGX~1\AppData\
Local\Temp\tmp9rcb_smb'
       cwd: C:\Users\zhangxian\AppData\Local\Temp\pip-install-ce6jw092\rasterio
  Complete output (2 lines):
  INFO:root:Building on Windows requires extra options to setup.py to locate needed GDAL files. More information is available in the README.
  ERROR: A GDAL API version must be specified. Provide a path to gdal-config using a GDAL_CONFIG environment variable or use a GDAL_VERSION environment variable.
  ----------------------------------------
ERROR: Command errored out with exit status 1: 'E:\SoftWare\Anaconda3\python.exe' 'E:\SoftWare\Anaconda3\lib\site-packages\pip\_vendor\pep517\_in_process.py' get_requires_for_buil
d_wheel 'C:\Users\ZHANGX~1\AppData\Local\Temp\tmp9rcb_smb' Check the logs for full command output.

Expected behavior
install successfully.

Screenshots

Additional context
Add any other context about the problem here.

Hi all, I've downloaded .zip project and then I imported it as a PyCharm project. If I try to execute the command with correct params and any type of command console returns me this error -->


I tried to install FFTW libraries but I got the same error

How can I install this great package without any error? Thanks in advance

Hi,
I've noticed that function _ehs uses np.histogram2d to calculate H used in ISSM, which isn't normalised, its just a sum of values.
https://github.com/up42/image-similarity-measures/blob/8328230be0ab45cf3c01db301db4064564298a5c/image_similarity_measures/quality_metrics.py#L165C1-L171C50
In original paper on ISSM authors use normalized joint histogram, the normalisation being number of pixels in image (MxN for image of dimension MxN).

During first usage of ssim function i am getting warning:

python3.9/site-packages/image_similarity_measures/quality_metrics.py:197: FutureWarning: `multichannel` is a deprecated argument name for `structural_similarity`. It will be removed in version 1.0. Please use `channel_axis` instead.
  return structural_similarity(org_img, pred_img, data_range=max_p, multichannel=True)

When the code calculates FSIM for multi-band data, the actual single-band results are averaged, and the calculations are all Nan-valued, even with my randomly generated matrices

I saw in a file that there was a TODO to improve performance, but any update on it? Comparing two simple png images using UIQ takes 2-3 minutes if not more, which is going to be quite annoying when dealing with as many images as i plan on doing.

Cool project though! Very nice to use otherwise

the following error occurs when I try to install:

ERROR: Could not find a version that satisfies the requirement image-similarity-measures (from versions: none)
ERROR: No matching distribution found for image-similarity-measures

It would be nice to use Poetry, which will avoid issues like this