visual_motif_removal

visual_motif_removal copied to clipboard

About the release of synthesized dataset

1

Dear Author, Thanks for your awesome work! Are you willing to release the synthesized dataset for training/testing in the paper? Best wishes, Xiaodong Cun.

vinthony

I can't get the pretrained mdel, can you share again?

4

The url in run_demo.py is not useful, can you share it again.

chaojie12131243

Model NOT working on watermark removal examples out of specified test data

1

Hi, thanks for sharing this code. I was impressed by the watermark removal results provided by this model. However, when I was trying this model (I didn't do any modifications,...

zijunwei

Size of feature map mismatches with the paper

2

In the paper, the size of the central feature map is 4x4x512  But [the code do not apply pooling on the last DownConv block](https://github.com/amirhertz/visual_motif_removal/blob/c7d158e3b448efb596a89ac869d154ca5c0816a8/networks/unet_deeper.py#L51), so that the...

jinyu121

Traceback (most recent call last): File "/content/visual_motif_removal/demo/run_demo.py", line 6, in from utils.train_utils import load_globals, init_folders ModuleNotFoundError: No module named 'utils'

ghost

The previous code turned all the non-black pixels on the RGBA motif into (255,255,255,255), i.e. white. The authors' intent was probably setting only the alpha coordinate to its maximal value...

ievron