I'd like to test some deep learning techniques to extract buildings footprint from aerial imagery. I've found many references related to this problem (here, or here), but only providing the model architecture and the process of learning.

I don't have easy access to hardware material in order to train a model. So I was looking for existing model weights (hdf5 or onnx files from state-of-the-art works) or pre-packaged script I could directly use as-is, just to make tests.

Do you know any repository or library with such model weights provided ?

I don't need to classify objects on the image, but only identify buildings (binary classification).


1 Answer 1


You can run .py and .ipynb if you have no hardware access via Google collab. That way allows you to train majority of NNs posted on github.

Also I found this pretrained network with utils for importing the NN as .ONNX: https://github.com/cuicaihao/aerial-image-segmentation/ So I run the export_onnx.py in Collab with adding opset_version=11 to torch.onnx.export args and got the .ONNX for the segmentation network.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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