vpanjeta / deep-object-removal Goto Github PK

Hey there,

I couldn't find an email so I'm writing here.
Would you be interested in a paid collaboration with my company? We are looking to develop an app that removes objects from images (similar to this: https://apps.apple.com/us/app/touchretouch/id373311252) using inpaint technique. We will take care of the entire app development, including the actual component for the masking, etc'. The user will use his fingers to choose the "removed object". But we need help to develop to object inpaint removal algorithm under the hood.  
Is it something you'll find interesting?

Roi

Hi Vishal,

I installed Deep-Object-Removal by following your doc and everything is ok.

But when I run "python3 main.py", I encountered the problem below.

Deep-Object-Removal % python main.py
RuntimeError: module compiled against API version 0xc but this version of Numpy is 0xb
ImportError: numpy.core.multiarray failed to import
ImportError: numpy.core.umath failed to import
ImportError: numpy.core.umath failed to import

This makes me a little confused.
Finally, I found the 0xc is for Numpy 1.14.1 and the 0xb is for Numpy 1.13.3.
My solution is to upgrade Numpy to 1.14.1 like below.
pip install -U numpy==1.14.1

And then, the issue has gone.

That's it.

Jarvis

Can you please specify how can we train the model with our own dataset?

Hi,
We have tried to maximise pen size to 200 from 20 so that we can use this program to delete text over images.
When we tried to fill a larger body than the max preset pen size we are getting black in the middle of the postprocessed area.
Is this because of a limiter inside the code that doesn’t allow drawn areas to be larger than the max pen size so we get black in the middle but not the perimeter of the drawn area.
Or is it because pre-trained model is trained in such way only small images and small areas to be deleted are used so once we try to draw larger areas to be deleted model doesn’t respond well?
Thank you for your support and creativity generating this code.

https://ibb.co/W2xcGkH

I am using the object removal model and it is working fine, but I don't know how to save the masked image, is there a way to do that?

I tried adding new Images to the project with the dimension 400x400 but unable to change it
Traceback (most recent call last):
File "main.py", line 76, in
sub_window = np.hstack((image, blank, filtered_image[:,:,[2, 1, 0]]))
File "/usr/local/lib/python3.6/dist-packages/numpy/core/shape_base.py", line 293, in hstack
return _nx.concatenate(arrs, 1)
ValueError: all the input array dimensions except for the concatenation axis must match exactly

Getting this error!