mirror/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2024-12-04 00:39:11 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Updated OE 32. Change OpenCV License to Apache 2 (markdown)

Vadim Pisarevsky 2020-06-06 17:46:27 +03:00
parent a0bee2f874
commit fc9287d5f3
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
OE-32.--Change-OpenCV-License-to-Apache-2.md

@ -16,14 +16,14 @@ BSD worked well for us for many years, but since the year 2000 computer vision w
* The technology is different, it's less of hand-made algorithms and more of industrial approaches and engineering.
* The market has grown significantly, with much more money and more players.
* Correspondingly, there is much more intellectual property (IP) there. Many companies, big and small, and many organizations work professionally in this area, create many new algorithms and patent many of them.
* Correspondingly, there is much more intellectual property (IP) there. Many and organizations companies, big and small, work professionally in this area, create lots of new algorithms and patent many of them.
* Software landscape has changed significantly. In 2000 OpenCV was one of very few computer vision libraries available. Now there are many solutions, especially in deep learning area, that are cross-platform, efficient, have rich functionality and strong communities behind them.
* OpenCV itself has changed significantly. It used to be a small C library of classical computer vision algorithms, solely developed by one team in one company. Now it contains 1,000,000+ lines of code and 70+ modules, and about half of the code is contributed by the community. Each working day about 5 patches are integrated into the library.
So, everything in computer vision nowadays, including OpenCV, is growing and developing fast, which is great!
On the other hand, such a rapid growth brings some potential and real issues. One issue is stability, which we mostly solved using our excellent and constantly improving Continuous Integration System and the large set of unit tests. Another issue is IP cleanness of the code. Which is as important as the reliability, especially for commercial products. That is, when people take OpenCV and use in their applications, they should have some confidence that OpenCV is robust, but also is clean. While OpenCV, just like almost any other software on a planet, cannot provide 100% guarantee of robustness and cleanness, we need to constantly work on it. As the queen in Alice in Wonderland said: _"My dear, here we must run as fast as we can, just to stay in place"_.
On the other hand, such a rapid growth brings some potential and real issues. One issue is stability, which we mostly solved using our excellent and constantly improving Continuous Integration System and the large set of unit tests. Another issue is IP cleanness of the code. Which is as important as the reliability, especially for commercial products. That is, when people take OpenCV and use in their applications, they should have some confidence that OpenCV is robust and that it is clean. While OpenCV, just like almost any other software on a planet, cannot provide 100% guarantee of robustness and cleanness, we need to constantly work on it. As the queen in Alice in Wonderland said: _"My dear, here we must run as fast as we can, just to stay in place"_.
There are 2 main cases when IP cleanness is violated:
There are 2 main cases when IP cleanness may be violated:
* Source code under incompatible license (specifically, GPL or "non-commercial use only") is integrated into OpenCV.
* Some patented algorithm is integrated into OpenCV.