Image segemantation for photography purposes.
Used the CelebA-HQ
image database to train Segnet using keras2.0 in jupyter notebook. Given a image expecting a segmentation
output like in this example (and hopefully better):
catalog.csv
- used to map images to their annotationsnotebooks
- folder containing training notebook(s) for jupytermodels
- current modelsscripts
- helper conversion scripts, mainly used to convertCelebA-HQ
database from text to csv which is parseable with pandas framework
- Great large-scale celebs photo images database: https://github.com/switchablenorms/CelebAMask-HQ
- https://github.com/namakemono/keras-segnet