lhyfst / dsfnet Goto Github PK

http://www.cbsr.ia.ac.cn/users/xiangyuzhu/projects/3DDFA/Code/3DDFA.zip
This link is broken, anyone can share the file with me? Thx

3DMM's shape_para, exp_para, and Tm values are very large, causing a large loss in the early stages of training. How can this be addressed?

Hi, incredible model, thank you for having released evaluation code, and my apologies if this is a dumb question, but has the training code been made available/open-sourced as well?

hi，I want to know how to generate the training data,too? Or is it possible for you to share training datasets?

Where to obtain “../SynergyNet/aflw2000_data/eval/ALFW2000-3D_pose_3ANG_excl.npy” and '../SynergyNet/aflw2000_data/eval/ALFW2000-3D_pose_3ANG_skip.npy'

I am really looking forward to checking out a demo for this repo. Is it possible to swap any reference of the BFM to the FLAME model, so it can be used for commercial purposes without having to pay for a $10k BFM license?

Hi， Greate work?

May I know when the code will be released?

Regards,

I want to visualize the created meshes through MeshLab from the out['face_uvm']. I need to change this UV map to a mesh using the provided code. But I couldn't find uv_triangles path in the code. What should be the path to it. Below is the code I am using:

Code

import numpy as np
import trimesh

def uvm2mesh(uv_position_map, only_foreface=True, is_extra_triangle=False):
    """
    If no texture map is provided, translate the position map to a point cloud
    :param uv_position_map:
    :param only_foreface:
    :param is_extra_triangle:
    :return:
    """
    uv_h, uv_w = UV_MAP_SIZE, UV_MAP_SIZE
    vertices = []
    colors = []
    triangles = []
    triangles = [[t[0] * uv_w + t[1], t[4] * uv_w + t[5], t[2] * uv_w + t[3], ] for t in uv_triangles]
    for i in range(uv_h):
        for j in range(uv_w):
            vertices.append(uv_position_map[i][j])
            colors.append([25, 25, 50, 128])

    vertices = np.array(vertices)
    triangles = np.array(triangles)
    colors = np.array(colors)
    face_mesh = trimesh.Trimesh(vertices=vertices, faces=triangles)

    return face_mesh