How to detect presence of object (highly occluded) in a scene?

There are specific features (small patterns, etc), which allow to say that object is present, but it is not enough for detection for YOLO or RPCNN.

How to detect small specific pattern in a whole image efficiently?

  • $\begingroup$ Do you have the entire image which is noisy? $\endgroup$ – Media Feb 26 '19 at 10:25
  • $\begingroup$ YOLO is still close to state-of-the-art for object detection, and there are plenty of hyperparameters to try, so it would be anyone's guess without more details of why it is failing for you. It may help if you show an image or two, with some details of what is detected (does it even have some small confidence of any object centred in the right grid location?), and what hyperparameters you have tried. $\endgroup$ – Neil Slater Feb 26 '19 at 11:07
  • $\begingroup$ One potential way you could go about doing this is a technique called masked projection, which has previously been used for 3D Face Recognition purposes where occlusions covering the facial surface were found to be an issue. You might find this paper of use: researchgate.net/publication/… $\endgroup$ – Michael Grogan Feb 26 '19 at 17:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.