CUVI + OpenCV

OpenCV is the single most widely used Open Source Image Processing/ Computer Vision library. You can use CUVI functions in conjunction with OpenCV which means if you have an application that already makes use of OpenCV, adding CUVI will not disturb the program flow i.e. you can replace OpenCV's Image Processing functions with CUVI's GPU accelerated version of the function and achieve speedups, at times, crossing over the boundary from non-realtime to realtime.

All CUVI functions accept pointer to the input and output image's data and related parameters as arguments which means you are free to read/write images any way you want to. This makes using CUVI fairly simple to use compared to some other libraries that require configuration and support limitedimage types.

The following example explains how a typical CUVI function can be called from an application that's already using OpenCV.

In this example, we read an image (Lena.pgm) from the filesystem using OpenCV’sHighGUI Image IO functions, create a window to display the input and processed image. The main part which is compute intensive is replaced by CUVI which adds acceleration. CUVI is first used to convert the input image from RGB to Gray scale and then we compute corners in the image using CUVI’s KLT corner detector which is ~12 times faster thenOpenCV’sSelectGoodFeatures. Experience the phenomenal speedup yourself by downloading CUVI now. You can download the source of this example from the link given at the end.