Custom object detection from new images dataset using Jetson Nano

I’d like to know if somebody have information about how to create a custom object detection using a new dataset of images, previously preprocessed and filtered. I have found this interesting site:

Also if somebody have used squares or polygons for image classification using Jetson Nano (labelimg or labelme), It will be usefull.



You can find an example to train an SSD-MobileNet with a custom dataset below.


1 Like