I have some images with anomalies, like small cracks, but it's an imbalanced dataset.
Please, suggest some effective scalable approaches. Should I consider convolutional auto-encoders? It's supposed to work in realtime, like video output.
$\begingroup$
$\endgroup$
-
$\begingroup$ Could you please suggest how many training samples you have? For these types of problems it's sometimes better to consider semi supervised learning. $\endgroup$ – pedrum Jul 27 '20 at 19:53
-
$\begingroup$ I have some training images with no cracks 1000 (majority) and with cracks 50. can you elaborate on semi-supervised approach? $\endgroup$ – ans_ak Jul 28 '20 at 4:40
-
$\begingroup$ I know that these may look like useless details to you, but, please, next time, put a little bit more effort into writing your posts. For example, do not write "i" rather than "I", and "it's" is different than "its". $\endgroup$ – nbro♦ Jul 28 '20 at 12:51
