Zed 2 + TensorRT + YoloV4 Combination

Hello, so I’m currently using an intel realsense camera to run object detection. My pipeline is that I grab an rgb image, feed it to the model, and it returns the bounding boxes.

I’m looking to switch to the zed2 camera which has a slightly different sdk. Is it possible to utilize their object detection with tensorRT? I know I could use the same process as I currently am, but their sdk takes care of the object detection, so it cuts out the middleman of setting up tensorRT and Cuda.

Thank you.