I have some data with ground truth that looks like a binary step function, where part of it is 0 and part is one. An example for the GT can be like 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0

or something like this

enter image description here

I have a hard time to come up with a loss function that can optimize this problem, the simplest option would be something like CrossEntropy or BinaryCrossEntropy, but I am wondering if there is any other loss that I try.

Something that can take into account the property that when the GT is one (1) it is continuous 1 and when it is zero it is continuous.

To give a little more information, for example, I will never have a GT that be like this 0 0 1 0 1 0 1 also I will never have a GT like this 0 0 0 0 1 1 1 0 0 0 0 1 1 1 0 0 0. In other words, I will have one time ones in a continuous way (it can be at the start or middle or end) but it wont be two discontinuous 1s. Is there any loss function that take into account this properties?


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.