Remove the global rotation of the generated model

[qianyunw]

Hi~ Thank you so much for your amazing work! I originally wanted to use the following code to remove the global rotation of the generated model, but because of the limited expressiveness of the parameters, the results are different from the original model. I would like to ask how to calculate directly on the vertex coordinates of the original model?

params = np.load(out_param, allow_pickle=True)
params = params[()]
#pose = np.hstack((params['pose'], np.zeros(15-params['pose'].shape[0])))
pose = np.zeros(15)

pose[6] = params['pose'][3]
pose[7] = params['pose'][4]
pose[8] = params['pose'][5]

expression = np.hstack((params['expression'], np.zeros(100-params['expression'].shape[0])))
shape = np.hstack((params['shape'], np.zeros(300-params['shape'].shape[0])))
flame_genral_model = load_model(config.flame_model_path)
generated_neutral_mesh = verts_decorated(ch.array([0.0,0.0,0.0]),
                                    ch.array(pose),
                                    ch.array(flame_genral_model.r),
                                    flame_genral_model.J_regressor,
                                    ch.array(flame_genral_model.weights),
                                    flame_genral_model.kintree_table,
                                    flame_genral_model.bs_style,
                                    flame_genral_model.f,
                                    bs_type=flame_genral_model.bs_type,
                                    posedirs=ch.array(flame_genral_model.posedirs),
                                    betas=ch.array(np.hstack((shape,expression))),#betas=ch.array(np.concatenate((theta[0,75:85], np.zeros(390)))), #
                                    shapedirs=ch.array(flame_genral_model.shapedirs),
                                    want_Jtr=True)
neutral_mesh = Mesh(v=generated_neutral_mesh.r, f=generated_neutral_mesh.f)
neutral_mesh.write_obj(out_mesh)
print("Saved neutral mesh file to " + out_mesh)

[qianyunw]

I also tried this, but it didn't work TAT :

params = np.load(out_param, allow_pickle=True)
params = params[()]
#pose = np.hstack((params['pose'], np.zeros(15-params['pose'].shape[0])))

pose = np.zeros(15)
pose[0] = params['pose'][0]
pose[1] = params['pose'][1]
pose[2] = params['pose'][2]
                
print(pose)
flame_genral_model = load_model(config.flame_model_path)
bs_type=flame_genral_model.bs_type
posedirs=ch.array(flame_genral_model.posedirs)
v_posed = vertices[0] - posedirs.dot(posemap(bs_type)(pose))

neutral_mesh = Mesh(v=v_posed, f=template_mesh.f)
neutral_mesh.write_obj(out_mesh)
print("Saved neutral mesh file to " + out_mesh)

[TimoBolkart]

What exactly do you want to do? Just output the RingNet prediction without the global rotation?

[HaoKun-Li]

It seems that some parameters in the "pose" will affect the open of the jaw and the rotation of the whole mesh. If we want to remove the global rotation but maintain the shape of the jaw, maybe we should keep the parameters about jaw rotation in the "pose" parameter. Then, set other pose parameters as zero.

Here is my code:

def make_prdicted_mesh_without_srt(params, flame_model_path):
    pose = np.zeros(15)
    # whole mesh rotation
    # pose[:3] = params['pose'][0:3]
    # jaw rotation
    pose[6:9] = params['pose'][3:]
    expression = np.hstack((params['expression'], np.zeros(100-params['expression'].shape[0])))
    shape = np.hstack((params['shape'], np.zeros(300-params['shape'].shape[0])))
    flame_genral_model = load_model(flame_model_path)
    generated_neutral_mesh = verts_decorated(ch.array([0.0,0.0,0.0]),
                        ch.array(pose),
                        ch.array(flame_genral_model.r),
                        flame_genral_model.J_regressor,
                        ch.array(flame_genral_model.weights),
                        flame_genral_model.kintree_table,
                        flame_genral_model.bs_style,
                        flame_genral_model.f,
                        bs_type=flame_genral_model.bs_type,
                        posedirs=ch.array(flame_genral_model.posedirs),
                        betas=ch.array(np.hstack((shape,expression))),#betas=ch.array(np.concatenate((theta[0,75:85], np.zeros(390)))), #
                        shapedirs=ch.array(flame_genral_model.shapedirs),
                        want_Jtr=True)
    neutral_mesh = Mesh(v=generated_neutral_mesh.r, f=generated_neutral_mesh.f)
    return neutral_mesh

predict mesh: predict mesh without whole rotation generate from my code: