Semantic photo manipulation with a generative image prior




  • Paper topic: Images
  • Software type: Online demo
  • Able to run a replicability test: False
  • Replicability score: 3
  • Software language:
  • License:
  • Build mechanism:
  • Dependencies:
  • Documentation score {0,1,2}: 0
  • Reviewer: Julie Digne <>
  • Time spent for the test (build->first run, timeout at 100min): 20min

Source code information


Misc. comments

An online demo is provided with test pictures. I was not able to delete doors and add domes in a convincing way, but adding grass and trees worked. Sometimes removing clouds only produces a blur. Tested on July 27th 2020.

If you want to contribute with another review, please follow these instructions.

Please consider to cut/paste/edit the raw JSON data attached to this paper.