File size: 11,258 Bytes
cd47a59 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 | import pip
# pip.main(["install", "pandas"]) # Unused and crashes with numpy alias
pip.main(["install", "tqdm"])
import bpy
import sys
import numpy as np
# Numpy 2.0 backward compatibility for Blender 3.6
sys.modules['numpy._core'] = sys.modules.get('numpy.core', np)
if 'numpy.core.multiarray' in sys.modules:
sys.modules['numpy._core.multiarray'] = sys.modules['numpy.core.multiarray']
# import pandas as pd # Unused
from pathlib import Path
import os
from ast import literal_eval
from tqdm import tqdm
from contextlib import contextmanager
import pathlib
this_script_path = pathlib.Path(__file__).parent.resolve()
W_FACE_AND_COLOR_FILE = this_script_path / "blend" / "smpl_mesh_info.npy"
FORMAT_LDR = "PNG"
COLOR_DEPTH_LDR = 8
SAMPLES = 1
COLOR_MODE = "RGB"
def setup_device(use_id):
bpy.context.scene.render.engine = "CYCLES"
bpy.context.preferences.addons["cycles"].preferences.compute_device_type = "CUDA"
bpy.context.scene.cycles.device = "GPU"
bpy.context.preferences.addons["cycles"].preferences.get_devices()
print(bpy.context.preferences.addons["cycles"].preferences.compute_device_type)
for i, device in enumerate(
bpy.context.preferences.addons["cycles"].preferences.devices
):
if i == use_id or "CPU" in device["name"]:
device["use"] = True # Using all devices, include GPU and CPU
else:
device["use"] = False # Using all devices, include GPU and CPU
print(device["name"], "USE:", bool(device["use"]))
class SingleDataset:
def __init__(self, smpl_folder, smpl_suffixes=["npy", "npz"]):
self.smpl_folder = Path(smpl_folder)
self.out_folder = Path(smpl_folder).parent
self.smpl_paths = []
self.bboxes = []
self.valid_index = []
folder = self.out_folder
self.smpl_paths = sorted(
[
path for i in smpl_suffixes
for path in (folder / "smpl_results").glob("*." + i)
]
)
self.output_paths = [self.out_folder for smpl_path in self.smpl_paths]
# Skip finished smpl_path. Enable it if want to continue processing only remaining imgs.
smpl_fns = [
os.path.splitext(os.path.basename(smpl_path))[0]
for smpl_path in self.smpl_paths
] # Example smpl_fn:0000
imgs_output_path = [
os.path.join(str(self.output_paths[i]), "visualized_imgs", f"{smpl_fn}.png")
for i, smpl_fn in enumerate(smpl_fns)
]
imgs_already_exist = [os.path.exists(smpl_path) for smpl_path in imgs_output_path]
imgs_index_to_inference = np.where(np.array(imgs_already_exist) == False)[0]
smpl_paths_copy = list(self.smpl_paths)
output_paths_copy = list(self.output_paths)
self.smpl_paths = [smpl_paths_copy[img_index] for img_index in imgs_index_to_inference]
self.output_paths = [output_paths_copy[img_index] for img_index in imgs_index_to_inference]
print(f"finish loading {imgs_index_to_inference.shape[0]} frames data, \
skip {len(smpl_fns) - imgs_index_to_inference.shape[0]} existing images")
def load_smpl(smpl_path):
return np.load(smpl_path, allow_pickle=True).item()
@contextmanager
def stdout_redirected(to=os.devnull):
"""
Redirects stdout to a specified file.
Usage:
with stdout_redirected(to=filename):
print("from Python")
os.system("echo non-Python applications are also supported")
"""
fd = sys.stdout.fileno()
# Save a copy of the original stdout file descriptor
original_stdout_fd = os.dup(fd)
# Redirect stdout to the specified file
with open(to, 'w') as file:
os.dup2(file.fileno(), fd)
try:
yield
finally:
# Restore the original stdout
os.dup2(original_stdout_fd, fd)
os.close(original_stdout_fd)
def rendering_pipeline(dataset, ref_img_path):
scene = bpy.context.scene
scene.render.image_settings.file_format = str(FORMAT_LDR)
scene.render.image_settings.color_depth = str(COLOR_DEPTH_LDR)
scene.render.image_settings.color_mode = str(COLOR_MODE)
scene.render.resolution_percentage = 100
scene.render.use_persistent_data = True
scene.cycles.use_denoising = False
camera = bpy.data.objects["Camera"]
camera.data.clip_start = 0.05
camera.data.clip_end = 1e12
camera.data.cycles.samples = SAMPLES
scene.cycles.samples = SAMPLES
bpy.context.scene.view_layers["ViewLayer"].use_pass_z = True
render_layers = bpy.context.scene.view_layers
if "mesh_collection" not in bpy.data.collections.keys():
mesh_collection = bpy.data.collections.new("mesh_collection")
bpy.context.scene.collection.children.link(mesh_collection)
mesh_collection = bpy.data.collections.get("mesh_collection")
mat_semantic = bpy.data.materials.get("Semantic")
mat_normal = bpy.data.materials.get("Normal")
result_dict = np.load(W_FACE_AND_COLOR_FILE, allow_pickle=True).item()
faces = result_dict["faces"]
verts_color = result_dict["verts_color"][:, ::-1].astype(np.float32) / 255.0
result_dict_list = []
processed = []
for path in tqdm(dataset.smpl_paths, total=len(dataset.smpl_paths), desc="Loading smpls into RAM"):
result = load_smpl(path)
processed.append(result)
for smpl in tqdm(processed, total=len(dataset.smpl_paths), desc="Loading smpls into RAM"):
result_dict_list.append(smpl)
for smpl_path, output_path, result_dict in tqdm(
zip(dataset.smpl_paths, dataset.output_paths, result_dict_list),
total=len(dataset.output_paths),
desc="Rendering Images",
miniters=10,
):
render_path = output_path
smpl_fn, _ = os.path.splitext(os.path.basename(smpl_path))
smpl_fn = smpl_fn.split(".")[0]
ori_img = bpy.data.images.load(ref_img_path)
ori_img.name = "ori_img.png"
ori_img.colorspace_settings.name = "Raw"
bpy.data.scenes["Scene"].node_tree.nodes["Image"].image = ori_img
cam_t = result_dict["cam_t"][0]
verts = result_dict["verts"][0] + cam_t
img_size = result_dict["render_res"].astype(np.int32)
camera.data.sensor_width = img_size.max()
camera.data.lens = result_dict["scaled_focal_length"]
scene.render.resolution_x = img_size[0]
scene.render.resolution_y = img_size[1]
# make mesh
new_mesh = bpy.data.meshes.new("smpl_mesh")
new_mesh.from_pydata(verts, edges=[], faces=faces)
# make object from mesh
new_object = bpy.data.objects.new("new_object", new_mesh)
# make collection
new_object.rotation_euler[0] = -np.pi / 2
for f in new_object.data.polygons:
f.use_smooth = True
# add object to scene collection
mesh_collection.objects.link(new_object)
# Render Normal Map and Depth Map
bpy.data.scenes["Scene"].node_tree.nodes["Depth Output"].base_path = (
os.path.join(render_path, "depth")
)
bpy.data.scenes["Scene"].node_tree.nodes["Visualize Output"].base_path = (
os.path.join(render_path, "visualized_imgs")
)
bpy.data.scenes["Scene"].node_tree.nodes["Mask Output"].base_path = (
os.path.join(render_path, "mask")
)
new_object.data.materials.append(mat_normal)
scene.view_settings.view_transform = "Raw"
bpy.data.worlds["World"].node_tree.nodes["Background"].inputs[0].default_value = (0, 0, 0, 1)
output_name = f"{smpl_fn}.png"
depth_path = os.path.join(render_path, "depth", output_name)
normal_path = os.path.join(render_path, "normal", output_name)
vis_path = os.path.join(render_path, "visualized_imgs", output_name)
mask_path = os.path.join(render_path, "mask", output_name)
semantic_path = os.path.join(render_path, "semantic_map", output_name)
scene.render.filepath = normal_path
for layer in render_layers:
# some condition
layer.use = layer.name == "ViewLayer"
bpy.context.scene.render.film_transparent = True
with stdout_redirected():
bpy.ops.render.render(write_still=True)
if os.path.isfile(depth_path):
os.remove(depth_path)
if os.path.isfile(os.path.join(render_path, "depth", f"{0:04d}.png")):
os.rename(os.path.join(render_path, "depth", f"{0:04d}.png"), depth_path)
if os.path.isfile(vis_path):
os.remove(vis_path)
if os.path.isfile(os.path.join(render_path, "visualized_imgs", f"{0:04d}.png")):
os.rename(
os.path.join(render_path, "visualized_imgs", f"{0:04d}.png"), vis_path
)
if os.path.isfile(mask_path):
os.remove(mask_path)
if os.path.isfile(os.path.join(render_path, "mask", f"{0:04d}.png")):
os.rename(os.path.join(render_path, "mask", f"{0:04d}.png"), mask_path)
# This is to reference the vertex color layer later
vertex_colors_name = "vert_colors"
# Here the color layer is made on the mesh
new_mesh.vertex_colors.new(name=vertex_colors_name)
color_layer = new_mesh.vertex_colors[vertex_colors_name]
# We loop over all the polygons
for poly in new_mesh.polygons:
# We get the polygon index and the corresponding mesh index
for vert_i_poly, vert_i_mesh in enumerate(poly.vertices):
# We get the loop index from the polygon index
vert_i_loop = poly.loop_indices[vert_i_poly]
# We set the color for the vertex
rgb = verts_color[vert_i_mesh].tolist()
rgb.append(1)
color_layer.data[vert_i_loop].color = rgb
del rgb
mat_semantic.node_tree.nodes["Color Attribute"].layer_name = "vert_colors"
new_object.data.materials[0] = mat_semantic
scene.view_settings.view_transform = "Standard"
scene.render.filepath = semantic_path
for layer in render_layers:
# some condition
layer.use = layer.name == "VertexColor Layer"
bpy.context.scene.render.film_transparent = False
with stdout_redirected():
bpy.ops.render.render(write_still=True)
bpy.data.objects.remove(new_object, do_unlink=True)
bpy.data.meshes.remove(new_mesh)
bpy.data.images.remove(ori_img)
if os.path.isfile(os.path.join(render_path, "visualized_imgs", f"{0:04d}.png")):
os.remove(os.path.join(render_path, "visualized_imgs", f"{0:04d}.png"))
del new_mesh
del ori_img
del new_object
del result_dict_list
if __name__ == "__main__":
import sys
argv = sys.argv
print(f"Rendering:")
try:
argv.index("--device")
except:
print("Use Only CPU for Rendering")
else:
setup_device(int(argv[argv.index("--device") + 1]))
smpl_folder = os.path.abspath(argv[argv.index("--driving_path") + 1])
ref_img_path = os.path.abspath(argv[argv.index("--reference_path") + 1])
rendering_pipeline(SingleDataset(smpl_folder), ref_img_path)
|