Hello,
For a project we do, we would like to automatically segment and label house images from Google Street View. I have seen some work on building facade semantic segmentation but it seemed much more oriented towards big city multi-story buildings where we usually do not see any roof.
Our application is more oriented towards residential 1-2 story houses in which we want to identify the roof regions in the images (see attached pictures).
If you could recommend appropriate papers or even better, open source libraries to do this, that would be great. The ideal would be NOT to have to manually segment, label and train our own CNN (or whatever the underlying technique) to achieve our goal.
Regards,
Bruno
- Badges
- Science topic
More Bruno Martin's questions See All
Similar questions and discussions