Real-Time Object Detection with YOLO v2 Using GPU Coder

Walk through an example of real-time object detection using YOLO v2 in MATLAB®. We start with a published example in MATLAB that explains how to train a YOLO v2 object detector and, using GPU Coder™, we generate optimized CUDA code.

We verify the generated code by compiling it into a MEX file using nvcc and we find the generated MEX to run at about 80 frames per second on the test video file.

Using the hardware support package for NVIDIA® GPUs, we deploy the generated code to the Jetson Xavier board as a standalone application.

Other Resources
Object Detection Using Deep Learning: https://bit.ly/2K2NJui
Object Detection Using YOLO v2 Deep Learning: https://bit.ly/2KCszmh
NVIDIA GPU Support from GPU Coder: https://bit.ly/2IsDAnp


No comments