Spaces:
Running
on
Zero
Running
on
Zero
| # | |
| # Copyright (C) 2023, Inria | |
| # GRAPHDECO research group, https://team.inria.fr/graphdeco | |
| # All rights reserved. | |
| # | |
| # This software is free for non-commercial, research and evaluation use | |
| # under the terms of the LICENSE.md file. | |
| # | |
| # For inquiries contact george.drettakis@inria.fr | |
| # | |
| import torch | |
| import math | |
| from diff_gaussian_rasterization import GaussianRasterizationSettings, GaussianRasterizer | |
| from scene.gaussian_model import GaussianModel | |
| from utils.sh_utils import eval_sh | |
| def render(viewpoint_camera, pc : GaussianModel, pipe, bg_color : torch.Tensor, kernel_size, scaling_modifier = 1.0, require_coord : bool = True, require_depth : bool = True): | |
| """ | |
| Render the scene. | |
| Background tensor (bg_color) must be on GPU! | |
| """ | |
| # Set up rasterization configuration | |
| tanfovx = math.tan(viewpoint_camera.FoVx * 0.5) | |
| tanfovy = math.tan(viewpoint_camera.FoVy * 0.5) | |
| screenspace_points = torch.zeros_like(pc.get_xyz, dtype=pc.get_xyz.dtype, requires_grad=True, device="cuda") + 0 | |
| try: | |
| screenspace_points.retain_grad() | |
| except: | |
| pass | |
| raster_settings = GaussianRasterizationSettings( | |
| image_height=int(viewpoint_camera.image_height), | |
| image_width=int(viewpoint_camera.image_width), | |
| tanfovx=tanfovx, | |
| tanfovy=tanfovy, | |
| kernel_size = kernel_size, | |
| bg=bg_color, | |
| scale_modifier=scaling_modifier, | |
| viewmatrix=viewpoint_camera.world_view_transform, | |
| projmatrix=viewpoint_camera.full_proj_transform, | |
| sh_degree=pc.active_sh_degree, | |
| campos=viewpoint_camera.camera_center, | |
| prefiltered=False, | |
| require_coord = require_coord, | |
| require_depth = require_depth, | |
| debug=pipe.debug | |
| ) | |
| rasterizer = GaussianRasterizer(raster_settings=raster_settings) | |
| means3D = pc.get_xyz | |
| means2D = screenspace_points | |
| # If precomputed 3d covariance is provided, use it. If not, then it will be computed from | |
| # scaling / rotation by the rasterizer. | |
| scales = None | |
| rotations = None | |
| cov3D_precomp = None | |
| scales, opacity = pc.get_scaling_n_opacity_with_3D_filter | |
| rotations = pc.get_rotation | |
| # If precomputed colors are provided, use them. Otherwise, if it is desired to precompute colors | |
| # from SHs in Python, do it. If not, then SH -> RGB conversion will be done by rasterizer. | |
| shs = pc.get_features | |
| colors_precomp = None | |
| rendered_image, radii, rendered_expected_coord, rendered_median_coord, rendered_expected_depth, rendered_median_depth, rendered_alpha, rendered_normal = rasterizer( | |
| means3D = means3D, | |
| means2D = means2D, | |
| shs = shs, | |
| colors_precomp = colors_precomp, | |
| opacities = opacity, | |
| scales = scales, | |
| rotations = rotations, | |
| cov3D_precomp = cov3D_precomp) | |
| # Those Gaussians that were frustum culled or had a radius of 0 were not visible. | |
| # They will be excluded from value updates used in the splitting criteria. | |
| return {"render": rendered_image, | |
| "mask": rendered_alpha, | |
| "expected_coord": rendered_expected_coord, | |
| "median_coord": rendered_median_coord, | |
| "expected_depth": rendered_expected_depth, | |
| "median_depth": rendered_median_depth, | |
| "viewspace_points": means2D, | |
| "visibility_filter" : radii > 0, | |
| "radii": radii, | |
| "normal":rendered_normal, | |
| } | |
| # integration is adopted from GOF for marching tetrahedra https://github.com/autonomousvision/gaussian-opacity-fields/blob/main/gaussian_renderer/__init__.py | |
| def integrate(points3D, viewpoint_camera, pc : GaussianModel, pipe, bg_color : torch.Tensor, kernel_size : float, scaling_modifier = 1.0, override_color = None): | |
| """ | |
| integrate Gaussians to the points, we also render the image for visual comparison. | |
| Background tensor (bg_color) must be on GPU! | |
| """ | |
| # Create zero tensor. We will use it to make pytorch return gradients of the 2D (screen-space) means | |
| screenspace_points = torch.zeros_like(pc.get_xyz, dtype=pc.get_xyz.dtype, requires_grad=True, device="cuda") + 0 | |
| try: | |
| screenspace_points.retain_grad() | |
| except: | |
| pass | |
| # Set up rasterization configuration | |
| tanfovx = math.tan(viewpoint_camera.FoVx * 0.5) | |
| tanfovy = math.tan(viewpoint_camera.FoVy * 0.5) | |
| raster_settings = GaussianRasterizationSettings( | |
| image_height=int(viewpoint_camera.image_height), | |
| image_width=int(viewpoint_camera.image_width), | |
| tanfovx=tanfovx, | |
| tanfovy=tanfovy, | |
| kernel_size = kernel_size, | |
| bg=bg_color, | |
| scale_modifier=scaling_modifier, | |
| viewmatrix=viewpoint_camera.world_view_transform, | |
| projmatrix=viewpoint_camera.full_proj_transform, | |
| sh_degree=pc.active_sh_degree, | |
| campos=viewpoint_camera.camera_center, | |
| prefiltered=False, | |
| debug=pipe.debug, | |
| require_depth = True, | |
| require_coord=True | |
| ) | |
| rasterizer = GaussianRasterizer(raster_settings=raster_settings) | |
| means3D = pc.get_xyz | |
| means2D = screenspace_points | |
| opacity = pc.get_opacity_with_3D_filter | |
| # If precomputed 3d covariance is provided, use it. If not, then it will be computed from | |
| # scaling / rotation by the rasterizer. | |
| scales = None | |
| rotations = None | |
| cov3D_precomp = None | |
| if pipe.compute_cov3D_python: | |
| cov3D_precomp = pc.get_covariance(scaling_modifier) | |
| else: | |
| scales = pc.get_scaling_with_3D_filter | |
| rotations = pc.get_rotation | |
| depth_plane_precomp = None | |
| # If precomputed colors are provided, use them. Otherwise, if it is desired to precompute colors | |
| # from SHs in Python, do it. If not, then SH -> RGB conversion will be done by rasterizer. | |
| shs = None | |
| colors_precomp = None | |
| if override_color is None: | |
| if pipe.convert_SHs_python: | |
| shs_view = pc.get_features.transpose(1, 2).view(-1, 3, (pc.max_sh_degree+1)**2) | |
| dir_pp = (pc.get_xyz - viewpoint_camera.camera_center.repeat(pc.get_features.shape[0], 1)) | |
| # # we local direction | |
| # cam_pos_local = view2gaussian_precomp[:, 3, :3] | |
| # cam_pos_local_scaled = cam_pos_local / scales | |
| # dir_pp = -cam_pos_local_scaled | |
| dir_pp_normalized = dir_pp/dir_pp.norm(dim=1, keepdim=True) | |
| sh2rgb = eval_sh(pc.active_sh_degree, shs_view, dir_pp_normalized) | |
| colors_precomp = torch.clamp_min(sh2rgb + 0.5, 0.0) | |
| else: | |
| shs = pc.get_features | |
| else: | |
| colors_precomp = override_color | |
| # Rasterize visible Gaussians to image, obtain their radii (on screen). | |
| rendered_image, alpha_integrated, color_integrated, point_coordinate, point_sdf, radii = rasterizer.integrate( | |
| points3D = points3D, | |
| means3D = means3D, | |
| means2D = means2D, | |
| shs = shs, | |
| colors_precomp = colors_precomp, | |
| opacities = opacity, | |
| scales = scales, | |
| rotations = rotations, | |
| cov3D_precomp = cov3D_precomp, | |
| view2gaussian_precomp=depth_plane_precomp) | |
| # Those Gaussians that were frustum culled or had a radius of 0 were not visible. | |
| # They will be excluded from value updates used in the splitting criteria. | |
| return {"render": rendered_image, | |
| "alpha_integrated": alpha_integrated, | |
| "color_integrated": color_integrated, | |
| "point_coordinate": point_coordinate, | |
| "point_sdf": point_sdf, | |
| "visibility_filter" : radii > 0, | |
| "radii": radii} |