I was looking at the Python Package Index (PyPi) and noticed 2 very similar packages: opencv-contrib-python and opencv-python and wondering what the difference was. I looked at them and they had the exact same description and version numbers.
There are several additional functionalities that are not available in the standard opencv-python package, like the SIFT algorithm. So, if you need to heavily work on images and computer vision and have no hardware restrictions, definitely install opencv-contrib-python.
EDIT: On my win10 computer I created a new conda environment and installed the latest opencv-contrib-python with pip. I see the same error message when I try to create the tracker in the new environment. There is no problem with opencv-contrib-python ver. 4.4.0.46.
Hi,
i tried the script by suggested in the first answer on a jetson xavier nx (newest jetpack i think) and it completed successfully. However it seems like opencv-contrib-python still does not work. I still get the error
The opencv-python packages come in multiple flavors and only one of them should be installed as explicitly stated in their documentation. So if an install already has opencv-contrib-python installed, a request to install opencv-python (not an uncommon dependency for other packages) should be ignored as its functionality (and more) is already provided by opencv-contrib-python. Currently, users can end up in an environment with both opencv-python and opencv-contrib-python installed, and depending on order of install all kinds of breakage.
If you still get an error message such as Could not find a version that satisfies the requirement opencv-contrib-python (from versions: ) No matching distribution found for opencv-contrib-python, try the alternative to use apt-get instead of pip:
Update 1: If you are using Python 2.7 and get an error message such as Could not find a version that satisfies the requirement opencv-contrib-python (from versions: ) No matching distribution found for opencv-contrib-python, a quick solution is to not use pip but apt-get instead:
If you already have some OpenCV distribution (such as opencv-python-headless, opencv-python, opencv-contrib-python or opencv-contrib-python-headless) installed in your Python environment, you can force Albumentations to use it by providing the --no-binary qudida,albumentations argument to pip, e.g.
For example, on macOS, if you installed OpenCV with Homebrew, the contrib module should be included, but if you installed opencv-python with pip, it might not be included. You need to install OpenCV with pip install opencv-contrib-python.
df19127ead