Spaces:
Running on Zero
Running on Zero
File size: 8,111 Bytes
7826761 | 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 | from __future__ import annotations
import cv2
import numpy as np
import torch
import torch.nn.functional as F
import nvdiffrast.torch as dr
from camera import Camera
from .mesh import compute_face_normal, vis_shading_mesh
iris_vertex_array = np.array([
3931,3932,3933,3935,3936,3937,3939,3940,3941,3943,3944,3945,3947,3948,
3949,3951,3952,3953,3955,3956,3957,3959,3960,3961,3963,3964,3965,3967,
3968,3969,3971,3972,3973,3975,3976,3977,3979,3980,3981,3983,3984,3985,
3987,3988,3989,3991,3992,3993,3995,3996,3997,3999,4000,4001,4003,4004,
4005,4007,4008,4009,4011,4012,4013,4015,4016,4017,4019,4020,4021,4023,
4024,4025,4027,4028,4029,4031,4032,4033,4035,4036,4037,4039,4040,4041,
4043,4044,4045,4047,4048,4049,4051,4052,4053,4054,4056,4057,4058,4477,
4478,4479,4481,4482,4483,4485,4486,4487,4489,4490,4491,4493,4494,4495,
4497,4498,4499,4501,4502,4503,4505,4506,4507,4509,4510,4511,4513,4514,
4515,4517,4518,4519,4521,4522,4523,4525,4526,4527,4529,4530,4531,4533,
4534,4535,4537,4538,4539,4541,4542,4543,4545,4546,4547,4549,4550,4551,
4553,4554,4555,4557,4558,4559,4561,4562,4563,4565,4566,4567,4569,4570,
4571,4573,4574,4575,4577,4578,4579,4581,4582,4583,4585,4586,4587,4589,
4590,4591,4593,4594,4595,4597,4598,4599,4600,4602,4603,4604
], dtype=np.int32)
class Visualizer:
def __init__(
self,
flame_assets_path: str = "weights/flame/flame_assets.npz",
device: str = "cuda",
) -> None:
self.device = torch.device(device)
self.glctx = dr.RasterizeCudaContext()
flame_assets = np.load(flame_assets_path)
self.vis_edges = np.asarray(flame_assets["vis_edges"], dtype=np.int32)
faces_t = torch.from_numpy(np.asarray(flame_assets["faces"]))
self.faces_int32 = faces_t.to(dtype=torch.int32, device=self.device)
self.faces_int64 = faces_t.to(dtype=torch.int64, device=self.device)
num_vertices = int(self.faces_int64.max().item()) + 1
base_vertex_colors = torch.ones((num_vertices, 3), dtype=torch.float32, device=self.device)
iris_idx = torch.from_numpy(iris_vertex_array).to(dtype=torch.long, device=self.device)
base_vertex_colors[iris_idx] = torch.tensor([0.0, 1.0, 0.0], dtype=torch.float32, device=self.device)
self.base_vertex_colors = base_vertex_colors.contiguous()
def vis_align(self, image: np.ndarray, vertex_coord: np.ndarray, line_width: int = 1) -> np.ndarray:
image_vis = image.copy()
vertex_xy = np.asarray(vertex_coord, dtype=np.int32)
for edge in self.vis_edges:
x0, y0, x1, y1 = vertex_xy[edge].reshape(-1).tolist()
cv2.line(image_vis, (x0, y0), (x1, y1), (0, 127, 255), line_width, lineType=cv2.LINE_AA)
return image_vis
def vis_contour(self, image: np.ndarray, vertex_coord: np.ndarray, line_width: int = 1) -> np.ndarray:
image_vis = image.copy()
h, w = image_vis.shape[:2]
vertices = torch.from_numpy(np.asarray(vertex_coord, dtype=np.float32)).to(self.device)
vertices_ndc = torch.zeros((vertices.shape[0], 4), dtype=torch.float32, device=self.device)
vertices_ndc[:, 0] = vertices[:, 0] / float(w) * 2.0 - 1.0
vertices_ndc[:, 1] = vertices[:, 1] / float(h) * 2.0 - 1.0
vertices_ndc[:, 2] = 0.5
vertices_ndc[:, 3] = 1.0
rast_out = dr.rasterize(
self.glctx,
vertices_ndc.unsqueeze(0),
self.faces_int32,
resolution=[h, w],
)[0].squeeze(0)
mask = (rast_out[..., 3:] > 0).to(dtype=torch.uint8)
mask = (mask * 255).cpu().numpy()
contours, _ = cv2.findContours(mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
if contours:
cnt = max(contours, key=cv2.contourArea)
cv2.drawContours(image_vis, [cnt], -1, (0, 255, 0), line_width)
return image_vis
def vis_flame_normal_overlay(
self,
camera: Camera,
image: np.ndarray,
vertices: np.ndarray | torch.Tensor,
alpha: float = 0.5,
return_normal_map: bool = False,
) -> np.ndarray | tuple[np.ndarray, np.ndarray]:
image_tf = torch.from_numpy(np.asarray(image, dtype=np.uint8)).to(dtype=torch.float32, device=self.device) / 255.0
if isinstance(vertices, np.ndarray):
vertices_tf = torch.from_numpy(vertices).to(dtype=torch.float32, device=self.device)
else:
vertices_tf = vertices.to(dtype=torch.float32, device=self.device)
normal_map, mask = vis_shading_mesh(self.glctx, camera, vertices_tf, self.faces_int64)
overlay = (image_tf * (1.0 - alpha) + normal_map * alpha) * mask + image_tf * (~mask)
normal_map = (normal_map * 255.0).to(dtype=torch.uint8, device="cpu").numpy()
overlay = (overlay * 255.0).to(dtype=torch.uint8, device="cpu").numpy()
overlay = np.ascontiguousarray(overlay)
normal_map = np.ascontiguousarray(normal_map)
if return_normal_map:
return overlay, normal_map
return overlay
def vis_flame_shading_overlay(
self,
camera: Camera,
image: np.ndarray,
vertices: np.ndarray | torch.Tensor,
alpha: float = 0.5,
return_shading: bool = False,
shading_background: str = "black",
) -> np.ndarray | tuple[np.ndarray, np.ndarray]:
image_tf = torch.from_numpy(np.asarray(image, dtype=np.uint8)).to(dtype=torch.float32, device=self.device) / 255.0
if isinstance(vertices, np.ndarray):
vertices_tf = torch.from_numpy(vertices).to(dtype=torch.float32, device=self.device)
else:
vertices_tf = vertices.to(dtype=torch.float32, device=self.device)
vertices_pad = F.pad(vertices_tf, [0, 1], value=1.0)
viewproj = camera.get_full_proj.to(device=self.device, dtype=torch.float32)
vertices_clip = torch.matmul(vertices_pad, viewproj.transpose(0, 1))
rast_out, _ = dr.rasterize(
self.glctx,
vertices_clip.unsqueeze(0),
self.faces_int32,
resolution=(camera.height, camera.width),
)
rast_squeezed = rast_out.squeeze(0)
face_id = rast_squeezed[..., 3:]
mask = face_id > 0
face_id = torch.clamp_min(face_id - 1, 0).long()
face_normals = compute_face_normal(vertices_tf, self.faces_int64)
normal_map_world = face_normals[face_id].squeeze(-2)
view_rot = camera.get_w2v[:3, :3].to(device=self.device, dtype=torch.float32)
normal_map_view = torch.matmul(normal_map_world, view_rot.transpose(0, 1))
normal_map_view = F.normalize(normal_map_view, dim=-1) * -1.0
light_dir = torch.tensor([0.0, 0.25, 1.0], dtype=torch.float32, device=self.device)
light_dir = F.normalize(light_dir, dim=0).view(1, 1, 3)
diffuse = torch.clamp_min(torch.sum(normal_map_view * light_dir, dim=-1, keepdim=True), 0.0)
shading_scalar = torch.clamp(0.15 + 0.85 * diffuse, 0.0, 1.0)
color_map, _ = dr.interpolate(self.base_vertex_colors, rast_out, self.faces_int32)
color_map = color_map.squeeze(0)
shading = color_map * shading_scalar
overlay = (image_tf * (1.0 - alpha) + shading * alpha) * mask + image_tf * (~mask)
if shading_background == "black":
shading = shading * mask
elif shading_background == "white":
shading = shading * mask + (~mask).to(dtype=shading.dtype)
else:
raise ValueError("shading_background must be 'black' or 'white'")
shading = (shading * 255.0).to(dtype=torch.uint8, device="cpu").numpy()
overlay = (overlay * 255.0).to(dtype=torch.uint8, device="cpu").numpy()
overlay = np.ascontiguousarray(overlay)
shading = np.ascontiguousarray(shading)
if return_shading:
return overlay, shading
return overlay
|