---
id: image-similarity-measures
version: "0.3.6"
license: MIT
license_treatment: permissive
maintenance: abandoned
---

# image-similarity-measures — Evaluation metrics to assess the similarity between two images.

License: permissive · Maintenance: abandoned · Downloads: 173.6K/mo

## What it is and what it does

Image Similarity Measures is a Python package that computes eight standard metrics to quantify how similar two images are. It wraps implementations of RMSE, PSNR, SSIM, FSIM, ISSM, SRE, SAM, and UIQ into a unified API, with both a Python function interface and a command-line tool. The package depends on numpy, scikit-image, opencv-python, and phasepack to handle image I/O and metric computation.

The package is mature and stable (marked Production/Stable), but is no longer actively maintained—the last release was in May 2023 with no commits since. It works with Python 3.8 through 3.11 and installs with low friction as a pure wheel. Optional dependencies (pyfftw for faster FSIM, rasterio for TIFF reading) can be added at install time.

Use it for:

- Evaluate image reconstruction quality in super-resolution or denoising projects by comparing output to reference images.
- Benchmark image compression algorithms by measuring similarity loss across different compression levels.
- Validate image processing pipelines in remote sensing or satellite imagery workflows.
- Assess quality of generated or predicted images in machine learning model evaluation.
- Compare image enhancement or restoration results quantitatively in research or production workflows.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Computes eight standard image similarity metrics (RMSE, PSNR, SSIM, FSIM, ISSM, SRE, SAM, UIQ) to quantify how similar two images are, via Python API or command-line tool.

Yes, if you need multiple standard image similarity metrics in one package and can accept no future maintenance. The package is stable, permissively licensed, and has low install friction. However, do not rely on it for critical production systems or expect bug fixes—consider it a snapshot tool. For active development or long-term support, evaluate maintained alternatives.

## Install

pip install image-similarity-measures
uv add image-similarity-measures
poetry add image-similarity-measures

## Installing image-similarity-measures

Before you install: Low friction install with a pure-Python wheel. Package is abandoned (last release 2023-05-04, no activity for 1198 days), so expect no maintenance or bug fixes going forward.

License in practice: MIT license is permissive; you can use this package freely in commercial and private projects with minimal restrictions.

Quickstart:

from image_similarity_measures.evaluate import evaluation

evaluation(org_img_path="original.tif", pred_img_path="predicted.tif", metrics=["rmse", "psnr"])

Requires Python >=3.8; images must be in channel-last format; depends on numpy, scikit-image, opencv-python, and phasepack.

Verify before relying:

- Whether the package works reliably with Python 3.10 and 3.11 given its abandoned status and lack of recent testing.
- Whether all eight metrics produce numerically correct results or if any have known issues that were never fixed.
- Performance characteristics of each metric on typical image sizes and whether optional speedups (pyfftw, rasterio) are still maintained.

## Package facts

- License: MIT (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: abandoned
- Downloads: 173.6K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

image similarity metrics, compare two images, image quality assessment, SSIM PSNR RMSE, image evaluation metrics, structural similarity index, image difference measurement, image-quality, metrics, abandoned

[View on SkillFed](https://skillfed.io/packages/image-similarity-measures) · [View on PyPI](https://pypi.org/project/image-similarity-measures/)