This project has a gitsubmodule to for seam-carving make sure to clone the submodule properly if you would like to resize your backgroud images
generateDataset.blend is an empty project to run the script in.
Populate the Backgrounds/ and Models/ directory with your own assets or use the provided example assets.
The models within the Models/ directory must have the same object name as the file itself.
The background images must all be the same size. 600 x 600. If they are not we provide the utility resize_images.py to make all the images the same size. If usint this utility make sure SHOULD_DOWNSIZE = False is set in the seam_carving.py file. This resizing method will take a long time to run.
Open generateDataset.blend in Blender
Run the generateDataset.py from Blender using the scripting tab in the generateDataset.blend project.
After running the script, see results in the output/ directory.
To visualize the results of the bounding boxes run display_image.py
https://github.com/andrewdcampbell/seam-carving
Image Credit: https://libreshot.com/wp-content/uploads/2019/12/foggy-road.jpg https://libreshot.com/wp-content/uploads/2018/12/road.jpg https://upload.wikimedia.org/wikipedia/commons/b/bb/Silaanyo_road%2C_habarjeclo_Road..jpg https://upload.wikimedia.org/wikipedia/commons/6/6b/Road_in_Norway.jpg
