423A35C7/pytorch3d
1
0
Fork 0
mirror of https://github.com/facebookresearch/pytorch3d.git synced 2025-08-02 20:02:49 +08:00
Code Issues Packages Projects Releases Wiki Activity

Add full-image PSNR metric

Browse Source 
Summary: Reports also the PSNR between the unmasked G.T. image and the render.

Reviewed By: bottler

Differential Revision: D38655943

fbshipit-source-id: 1603a2d02116ea1ce037e5530abe1afc65a2ba93
This commit is contained in:
David Novotny 2022-08-12 07:56:44 -07:00 committed by Facebook GitHub Bot
parent a91f15f24e
commit 7b985702bb
2 changed files with 26 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
pytorch3d/implicitron/evaluation/evaluate_new_view_synthesis.py
View File

@ -262,6 +262,10 @@ def eval_batch(
else torch.ones_like(mask_fg) else torch.ones_like(mask_fg)
) )
# unmasked g.t. image
image_rgb = frame_data.image_rgb
# fg-masked g.t. image
image_rgb_masked = mask_background( image_rgb_masked = mask_background(
# pyre-fixme[6]: Expected `Tensor` for 1st param but got # pyre-fixme[6]: Expected `Tensor` for 1st param but got
# `Optional[torch.Tensor]`. # `Optional[torch.Tensor]`.
@ -330,12 +334,26 @@ def eval_batch(
if break_after_visualising: if break_after_visualising:
breakpoint() # noqa: B601 breakpoint() # noqa: B601
# add the rgb metrics between the render and the unmasked image
for rgb_metric_name, rgb_metric_fun in zip(
("psnr_full_image", "rgb_l1_full_image"), (calc_psnr, rgb_l1)
):
results[rgb_metric_name] = rgb_metric_fun(
image_render,
image_rgb,
mask=mask_crop,
)
if lpips_model is not None: if lpips_model is not None:
im1, im2 = [ for gt_image_type in ("_full_image", ""):
2.0 * im.clamp(0.0, 1.0) - 1.0 im1, im2 = [
for im in (image_rgb_masked, cloned_render["image_render"]) 2.0 * im.clamp(0.0, 1.0) - 1.0 # pyre-ignore[16]
] for im in (
results["lpips"] = lpips_model.forward(im1, im2).item() image_rgb_masked if gt_image_type == "" else image_rgb,
cloned_render["image_render"],
)
]
results["lpips" + gt_image_type] = lpips_model.forward(im1, im2).item()
# convert all metrics to floats # convert all metrics to floats
results = {k: float(v) for k, v in results.items()} results = {k: float(v) for k, v in results.items()}

3
tests/implicitron/test_evaluation.py
View File

@ -256,10 +256,13 @@ class TestEvaluation(unittest.TestCase):
lower_better = { lower_better = {
"psnr": False, "psnr": False,
"psnr_fg": False, "psnr_fg": False,
"psnr_full_image": False,
"depth_abs_fg": True, "depth_abs_fg": True,
"iou": False, "iou": False,
"rgb_l1": True, "rgb_l1": True,
"rgb_l1_fg": True, "rgb_l1_fg": True,
"lpips": True,
"lpips_full_image": True,
} }
for metric in lower_better: for metric in lower_better: