Framed Outlay instead of color filled box

I’m new to using the Nvidia Jetson Nano. I traversed though the Nvidia self paced online training from “Setting up Jetson with Jetpack” through “Running the Live Camera Detection Demo”. I’d like to get more sophisticated, and attempting to make my first modification but clueless of where and how to attack the change.

I’d like to change the image outlay from a filled in colored box to a framed box with a clear interior. I searched the code of detectNet.ccp and see a variable OUTPUT_BBOX set to 1 and a comment about Caffe. Is this where one would start? I can’t find any documentation on the values of this variable anywhere. Where would I look or should I look elsewhere. Can you suggest a way to make the modification and an Nvidia class that could teach me more about modifying the output image?

I code in C++.

Environment info:
NVIDIA Jetson Nano (Developer Kit Version)
L4T 32.5.1 [ JetPack 4.5.1 ]
Ubuntu 18.04.5 LTS|
Kernel Version: 4.9.201-tegra|
CUDA 10.2.89|
CUDA Architecture: 5.3|
OpenCV version: 4.1.1
OpenCV Cuda: NO|
Vision Works:
VPI: ii libnvvpi1 1.0.15 arm64 NVIDIA Vision Programming Interface library
Vulcan: 1.2.70

Hi @steven.sparks, the detection overlay rendering code is found here:

That CUDA kernel draws one box on top of the image. You could attempt to modify it so that if the x/y coordinate it outside the range of your desired border, then it doesn’t fill the pixel. If you make code changes, remember to re-run make && sudo make install

Alternatively, you can see here for more general CUDA image manipulation routines:

If it’s easier, you can use openCV/cv2 to make custom rendering if you want.

Dusty, Thank you. I will experiment and let everyone see my coding results.