CUVI version 0.5 is cooked in our labs and we are doing testing and documentation at the moment. The new release will be out anytime in the coming week. We have been working for almost six months on the new framework that couldn’t get any simpler and easy to use. In this release we are also enabling our premium feature detectors that are 10 times faster than OpenCV 2.2
KLT: Kanade-Lucas-Tomasi Feature detector
KLT feature and corner detection is widely used in computer vision application in numerous scientific and industry applications. Our implementation of KLT is as good as OpenCV’s but only 10 times faster. See the result and benchmark with here. The KLT function in CUVI with minimum arguments is as simple as this:
CuviStatus klt(CuviImage* inputImage, cuviPointValue2D*& outputCorners, int *maxCorners)
We also provide additional parameters to fine tune your results. These parameters are Quality Level; that has the ability to prune out weak features, blockSizeforEigen that calculates the Individual Eigen Values in the given block neighborhood and the minNeighbourhoodBlock that selects only 1 feature/corner in that neighborhood.
CuviStatus klt(..., float qualityLevel, int blockSizeforEigen, int minNeighbourhoodBlock)
Harris Corner Detector
We provide two implementations of Harris Corner Detector in CUVI v0.5, one is the original and the other is by Peter Kovesi that you can easily toggle with a bool variable in the function. Both the implementations have their own pros and cons that you yourself can judge better. Our implementation of Harris is as fast as our KLT giving up to 10 times faster results than OpenCV’s CPU implementation.
Besides feature detection we have improved our Optical Flow implementations and introduced 2D-filters and color operations as well. Our framework works seamlessly with OpenCV, see the demo here. To sum up CUVI v0.5 will have:
- Feature Detection Algorithms, link
- Improved Optical Flow Implementations, link
- 2D Filters, link
- Color Operations, link
- Linux and Mac Support added
- And a new, better framework with OpenCV Interoperability
If you’re already a CUVI user we’ll email you as soon as we release v0.5, otherwise keep and eye on download page and join us on social media using the links below.