search

yfeng95 / face3d

Python tools for 3D face: 3DMM, Mesh processing(transform, camera, light, render), 3D face representations.
2.65k stars 610 forks source link

about the input of fit.fit_point #36

Open woshinierbaba opened 5 years ago

woshinierbaba commented 5 years ago

hello professor, i don't know it exactly what is the two input: x: (n, 2) image points X_ind: (n,) corresponding Model vertex indices is x just a picture? and is X_ind generated by morphable.generate_vertices?

kekedan commented 5 years ago

hello professor, i don't know it exactly what is the two input: x: (n, 2) image points X_ind: (n,) corresponding Model vertex indices is x just a picture? and is X_ind generated by morphable.generate_vertices?

the ame problem .do you got it?

zhongtao93 commented 5 years ago

the same question, can anyone have explantion?

tjusxh commented 5 years ago

I guess the x should be the 68 2d points. But I get 68 points from 2d image. But I get non-ideal result. How to normalization the points? generated_image00002

fitted_image00002

I resolve the question by x= x - np.mean(x,axis=0) x[:1] = -x[:,1] . But I don't sure if it is right.

zhongtao93 commented 5 years ago

I have get a good result. image image I use the code bellow, x is 68d face keypoint x = x / np.max(x) * 255 x -= np.mean(x)

Zenway-Wong commented 2 years ago

I have get a good result. image image I use the code bellow, x is 68d face keypoint x = x / np.max(x) * 255 x -= np.mean(x)

Hi, can you please share how you get the 2D point? And how to get fitted texture?

pfeducode commented 2 years ago

Have you solved this problem? Why is my output image reversed? The generated picture is as follows generated the orignial pose image aflw-test

and I use 68 key points to fit the generated image, and the fit picture is as follows，I dont know why i get such a strange picture fitted