0
$\begingroup$

After reading the state of the Art of object detection using the CNN's(R-CNN,Faster R-CNN,YOLO,YOLOv2,SSD) I was wondering if there is an efficient method that use deep learning with reinforcement learning for object detection ?

$\endgroup$
  • 1
    $\begingroup$ That's a bit vague/open-ended, do you have anything in mind? For instance, do you have any environment or problem where you are hoping RL will apply? RL needs a specific framing of a problem, it isn't an architecture choice such as R-CNN, YOLO etc. $\endgroup$ – Neil Slater Feb 23 '18 at 19:27
  • $\begingroup$ I meant they use RL for example instead of Region proposal or something similar $\endgroup$ – hbdz Feb 23 '18 at 19:36
2
$\begingroup$

To my extent of my knowledge, RL is used as a model for attention mechanism in object detection field (particularly the REINFORCE algorithm which is a 'flavor' of policy gradient methods). You can take a look at one of the papers that first proposed the method: RAM

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.