try to render mesh use PerspectiveCameras,but get nothing

[zhangzscn]

hello, try to render mesh by PerspectiveCameras,but the result is same color.
this is my code:

import torch
import numpy as np

import matplotlib.pyplot as plt

from pytorch3d.io import load_obj
from pytorch3d.structures import Meshes

from pytorch3d.renderer import (
PerspectiveCameras,
RasterizationSettings, MeshRenderer, MeshRasterizer, BlendParams,
SoftSilhouetteShader
)

if torch.cuda.is_available():
device = torch.device("cuda:0")
torch.cuda.set_device(device)
else:
device = torch.device("cpu")

verts, faces_idx, _ = load_obj("./data/smpl1.obj")
faces = faces_idx.verts_idx
verts = verts * 1000
smpl_mesh = Meshes(verts=[verts.to(device)], faces=[faces.to(device)])

focal_length = torch.tensor([[577.137, 577.137]], device=device)
principal_point = torch.tensor([[321.874, 243.694]], device=device)
R = torch.tensor([[[1, 0, 0], [0, 1, 0], [0, 0, 1]]], device=device)
T = torch.tensor([[0, 0, 0]], device=device)

cameras = PerspectiveCameras(focal_length=focal_length,
principal_point=principal_point,
R=R,
T=T,
device=device
)

blend_params = BlendParams(sigma=1e-4, gamma=1e-4)
raster_settings = RasterizationSettings(
image_size=[480, 640],
blur_radius=np.log(1. / 1e-4 - 1.) * blend_params.sigma,
faces_per_pixel=100,
)

silhouette_renderer = MeshRenderer(
rasterizer=MeshRasterizer(
cameras=cameras,
raster_settings=raster_settings
),
shader=SoftSilhouetteShader(blend_params=blend_params)
)

silhouette = silhouette_renderer(meshes_world=smpl_mesh)
silhouette = silhouette.cpu().numpy()

skt = silhouette.squeeze()[..., 3]
plt.figure(figsize=(10, 10))
plt.imshow(silhouette.squeeze()[..., 3])
plt.grid(False)
plt.show()

some of points in mesh like this:

v 0.19095382 -0.63784724 2.36624336
v 0.18782929 -0.62453818 2.35648489
v 0.19752663 -0.62003261 2.36658001
v 0.20136854 -0.62961197 2.37770987
v 0.19561735 -0.61190212 2.35865927
v 0.18636529 -0.61382329 2.35095119
v 0.20367119 -0.61277306 2.37667370
v 0.20711969 -0.61854726 2.38839769
v 0.21181732 -0.57649171 2.38347101
v 0.21480688 -0.57056987 2.39325476
v 0.21359174 -0.58262002 2.39287210
v 0.20959769 -0.58547872 2.38358545
v 0.21206786 -0.55020374 2.38399220
v 0.20713890 -0.54289997 2.37437510
v 0.20941487 -0.53281802 2.38327074
v 0.21310993 -0.53974956 2.39433289
v 0.17692581 -0.55943173 2.34842968
v 0.18259536 -0.55283284 2.34954000
v 0.18584865 -0.55930161 2.35169315
v 0.17977718 -0.56495500 2.35053730
v 0.16799158 -0.57012534 2.34174085
v 0.16867708 -0.57448936 2.34582734
v 0.16423194 -0.57906401 2.34025574
v 0.16457398 -0.57396770 2.33498240
v 0.17229365 -0.55443263 2.34643221
v 0.17860942 -0.54718447 2.34760237
v 0.17212476 -0.56533957 2.34667325
v 0.17059043 -0.56155694 2.34336567
v 0.16987452 -0.55765593 2.33099699
v 0.17081188 -0.56217968 2.33318520
v 0.16834566 -0.56386817 2.32591319
v 0.16717462 -0.55851430 2.32592106
v 0.17156522 -0.55517679 2.34110594
v 0.17206945 -0.55887848 2.33848882
v 0.17030457 -0.55547982 2.33541369
v 0.16870804 -0.55250055 2.33834958
v 0.16814694 -0.56643689 2.33782148
v 0.16593423 -0.56943667 2.33039045
v 0.16287780 -0.57174528 2.32257652
v 0.16481996 -0.56418616 2.31812334
v 0.19343680 -0.53859717 2.35593534
v 0.18853624 -0.54589862 2.35186768

[gkioxari]

From a first read, you camera definition seems off. You define the camera in screen space but don't set it that way.
If you need help, I'd advise to move this issue to Discussions and also read our notes more carefully

https://github.com/facebookresearch/pytorch3d/blob/main/docs/notes/cameras.md

[pramishp]

I also have a similar issue. any solutions?

[nguyenhoangvudtm23]

did this have solutions?