Image classification training process through DIGITS, and architecture.

I’d like to know the procedure of how image classification training can be happened using DIGITS, and the architecture of the training and object detection.

Yes, this is one of the primary use cases for digits. There is a great deal of information on line concerning this if you want to search for it.

https://github.com/NVIDIA/DIGITS/blob/master/docs/GettingStarted.md
https://developer.nvidia.com/digits-course-register