I am using visionworks library for calculating the disparity map using SGM algorithm. I wanted to know if there is any way to obtain the confidence score(or the cost ) of the disparity value predicted for every pixel int the image, using which I can improve the final disparity obtained by interpolation or some fusion technique.
Instead of high-level API, you can get more information in our low-level API.
For example, in llsgm, aggregated_cost may be what you want.
Thanks for your quick reply. Can you share link for the documentation of this function/API. I couldn’t find any.
Semi-Global Matching document can be found at:
NVIDIA Extension API
Vision Primitives API
The source code of our low-level SGBM can be found here:
Could you please tell me more specifically about where these relative document is? website, link? or kind of package? or where i can download these files
I has installed visionworks by installing Jetpack3.2, but i can not find the SGM documents you provided in my Jetson TX2
Here is the information of VisionWorks packages:
Package Name : libvisionworks Installed location: /usr/lib Description : Main package with pre-built shared libraries. ________________________________________________________________________________ Package Name : libvisionworks-dev Installed location: /usr/include /usr/lib /usr/lib/pkgconfig /usr/share/visionworks/cmake Description : Development package with headers, supplementary CMake and package config files. ________________________________________________________________________________ Package Name : libvisionworks-samples Installed location: /usr/share/visionworks/sources Description : Source code for samples, demos, and NVXIO. ________________________________________________________________________________ Package Name : libvisionworks-docs Installed location: /usr/share/visionworks/docs Description : Documentation package for this release of VisionWorks. ________________________________________________________________________________
You can find this information here: