I wish to process frames from camera using CUDA kernel and then display the result on screen. Could anyone suggest the best way to do this maybe without hurting in terms of cycle consumption ? Is there any example which uses frames from camera and displays the cuda processed result on screen which I can follow ?
Based on recommendation from folks on forum we are using NvEGLRenderer to display the incoming frames. However the render API used only takes “fd” as argument. Hence I am in a fix on how to proceed forward with this.
Please help me out.