[Multimedia] GPU processing of V4L2 live video on NVIDIA Jetson with CUDA

Graphics Processing Unit (GPU) image processing is fast and power-efficient. Getting video data into and out of the GPU can be slow and complicated. I will help you over those barriers. From this talk, you will learn how to move video data from a Video for Linux (V4L2) camera into the GPU on an NVIDIA Jetson system-on-chip. You will see how to process that data on the GPU using CUDA. I will also demonstrate how to use EGL to output directly to the display. These building blocks will help you implement your own software for processing images and displaying results in embedded devices such as edge-AI processors. No trip to the cloud required! Technology permitting, this talk will include a live demonstration showing raw Bayer video on a monitor. Follow Live Embedded Event on : Twitter :   / liveembedded   LinkedIn :   / live-embedded-078abb1b9