Hello,
I am going to use RL for image processing.
As far as I know, in order to apply RL in image processing, I need to convert features to states. Is it right to use LSTM? I think I need to use CNN to extract features and then feed those features to LSTM to create an environment for using RL. Is it right?