tink/nerfstudio
4
0
Fork 0
mirror of https://github.com/nerfstudio-project/nerfstudio.git synced 2026-11-11 18:44:21 +08:00
Code

Merge branch 'main' into ethan/visiblity-depth-nerfacto

Browse Source
This commit is contained in:
Matias Turkulainen 2024-03-10 19:10:27 +02:00 committed by GitHub
commit fa9d7ab45e
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 8 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
nerfstudio/data/datamanagers/full_images_datamanager.py
View File

@ -387,6 +387,8 @@ def _undistort_image(
mask = cv2.undistort(mask, K, distortion_params, None, newK) # type: ignore
mask = mask[y : y + h, x : x + w]
mask = torch.from_numpy(mask).bool()
if len(mask.shape) == 2:
mask = mask[:, :, None]
K = newK
elif camera.camera_type.item() == CameraType.FISHEYE.value:
@ -406,6 +408,8 @@ def _undistort_image(
mask = mask.astype(np.uint8) * 255
mask = cv2.fisheye.undistortImage(mask, K, distortion_params, None, newK)
mask = torch.from_numpy(mask).bool()
if len(mask.shape) == 2:
mask = mask[:, :, None]
K = newK
elif camera.camera_type.item() == CameraType.FISHEYE624.value:
fisheye624_params = torch.cat(
@ -498,6 +502,8 @@ def _undistort_image(
)
/ 255.0
).bool()[..., None]
if len(mask.shape) == 2:
mask = mask[:, :, None]
assert mask.shape == (undist_h, undist_w, 1)
K = undist_K.numpy()
else:

2
nerfstudio/models/instant_ngp.py
View File

@ -223,7 +223,7 @@ class NGPModel(Model):
return metrics_dict
def get_loss_dict(self, outputs, batch, metrics_dict=None):
image = batch["image"][..., :3].to(self.device)
image = batch["image"].to(self.device)
pred_rgb, image = self.renderer_rgb.blend_background_for_loss_computation(
pred_image=outputs["rgb"],
pred_accumulation=outputs["accumulation"],

2
nerfstudio/models/tensorf.py
View File

@ -308,7 +308,7 @@ class TensoRFModel(Model):
def get_loss_dict(self, outputs, batch, metrics_dict=None) -> Dict[str, torch.Tensor]:
# Scaling metrics by coefficients to create the losses.
device = outputs["rgb"].device
image = batch["image"][..., :3].to(device)
image = batch["image"].to(device)
pred_image, image = self.renderer_rgb.blend_background_for_loss_computation(
pred_image=outputs["rgb"],
pred_accumulation=outputs["accumulation"],