Object detection using the YOLO neural network and CUDA hardware and software architecture https://youtu.be/ZmsavPCiwLU
An experiment was conducted on the use of CUDA and tensor kernels in problems of object detection and recognition. This example uses a pre-trained object detection network such as YOLO v3 trained on the COCO dataset. Requires NVIDIA® GPU with CUDA support and compatible driver. I used Emgu.CV.runtime.windows.cuda version 4.4.0.4099 and all other Emgu.CV packages version 4.4.0.4099 (versions must match).