Pantelis Kaniouras

In case you are interested in road detection methods: https://github.com/ntelo007/road_detection_mtl

I will reverse the one hot encoding and then try to evaluate my images because if I don't, I receive really high accuracy without this being the truth. I will...

For starters, thank you for your immediate response. I am talking about Multiple tasks --> Semantics Segmentation + Instance Segmentation + Orientation Learning for example. Every task will have its...

Also, could you please tell me how I can see the details of the model variable that you proposed? I would like to be able to read the differences when...

> ResNet is a style of architecture that is made up of Residual blocks or modules, which has skip layers or connections to pass information from earlier in the network...

Let's assume that I have three classes. Background, class A and class B. Basically, I need to identify every pixel that belongs to class A and class B and the...

> Yes. You need to one-hot-encode them to work with Keras regardless of the loss function you use. To make things easy, you should have your masks in the format...

Same here

Same here

Thank you very much, that would be great!