imreadeval is a package to evaluate the performance of various python imread functions from different packages.
pip install imreadeval
PS. If you use anaconda, please install dependencies before pip install. The default dependency install is pip based.
Or just clone this repo, if you want to use updated version.
- Pillow https://github.com/python-pillow/Pillow
- OpenCV https://github.com/opencv/opencv
- Matplotlib https://github.com/matplotlib/matplotlib
- imageio (descendent of
scipy.io.imread, which is deprecated) https://github.com/imageio/imageio - SciKit-Image (depend on SciPy) https://github.com/scikit-image/scikit-image
- imread https://github.com/luispedro/imread
Just evaluate import imreadeval, then all imread packages installed will be tested, and the fastest one shows.
In [1]: import imreadeval INFO:root:
imreadperformance test: INFO:root: Pillow time: 0.8008584999999995. INFO:root: OpenCV is not installed. INFO:root: Matplotlib time: 0.8844109000000007. INFO:root: imageio time: 0.9442649999999997. INFO:root: SciKit-Image time: 0.9424957000000003. INFO:root: imread is not installed. INFO:root: Pillow is the fastest.
Or just import a fastest imread function by from imreadeval import imread.
In [1]: import imreadeval INFO:root:
imreadperformance test: INFO:root: Pillow time: 0.8008584999999995. INFO:root: OpenCV is not installed. INFO:root: Matplotlib time: 0.8844109000000007. INFO:root: imageio time: 0.9442649999999997. INFO:root: SciKit-Image time: 0.9424957000000003. INFO:root: imread is not installed. INFO:root: Pillow is the fastest.In [2]: x = imread('imreadeval/images/clouds.png')
or try to evaluate different imread functions using YOUR OWN IMAGES by
from imreadeval import imread_eval
optimum_package_name = imread_eval(filenames = ['filename1.jpg', 'filename2.png'], times = 1000)
git clonethe repo[email protected]:quxiaofeng/imreadeval.gitpip install -e .python -c "import imreadeval"
This package is done in Tsinghua University, Shenzhen International Graduate School.