Can TLT Detecnet_v2 be trained with thermal images?

Accelerated Computing Intelligent Video Analytics TAO Toolkit
1

Hi,

I wonder if I can train a model with thermal images on TLT/DetectNet_v2 (backbone as resnet18) and that model can work on any Jetson devices (especially on TX2.)

If it is possible, is there anything that i have to know before to start training. For example, is there any must to have regularizations that i have to do for the spec files to train with thermal images.

Thanks …

5

It is expected to work.
See https://ngc.nvidia.com/catalog/models/nvidia:tlt_facedetectir and Using FaceDetectIR model in Triton Server - #3 by Morganh