ramgendeploy / headswap-machine Goto Github PK

Hi Ramgen,
Thanks for your hard work.

I'm very new in this field.
I don't wanted to open new issue but didn't find right way to contact you. By the way I have followed you on twitter just now. I tried your code and found that "test_result.jpg" is showing dark face where as "source_1.jpg" is on fair side. Let me know if something needs to be changed. Attaching all files for your reference.

Hi Ramgen,
Thanks for your hard work.

I'm very new in this field.
I would like to test your code, but the pretrained model path not found:

pretrained_W = "/home/rama/legendfacesPipeline/models/bisenet_pretrained.pth"

Where can I download this model or which models can I use the same with "bisenet_pretrained.pth"?

Thanks!

Hi Ramgen,

it's me again ✌️

I'm have been working on your project for a days, but no luck to get the perfect image result.
I can not move the source mask and the crop face into place to make the composition with the headless template.

Maybe the problems is in function : "transformation_from_points":

def transformation_from_points(points1, points2):
points1 = points1.astype(np.float32)
points2 = points2.astype(np.float32)

c1 = np.mean(points1, axis=0)
c2 = np.mean(points2, axis=0)
points1 -= c1
points2 -= c2

s1 = np.std(points1)
s2 = np.std(points2)
points1 /= s1
points2 /= s2

U, S, Vt = np.linalg.svd(points1.T * points2)
R = (U * Vt).T

return np.vstack([np.hstack(((s2 / s1) * R,
                             c2.T - (s2 / s1) * R * c1.T)),
                  np.matrix([0., 0., 1.])])

Could you please explain for me how it work and how to get the perfect result?
Thank you!

is there is any way to create the headless template image..