PR-ENDO/render.py at main · SanoScience/PR-ENDO · GitHub

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
import torch
from scene import Scene
import os
import time
from tqdm import tqdm
from os import makedirs
from gaussian_renderer import render
import torchvision
from utils.general_utils import safe_state
from argparse import ArgumentParser
from arguments import ModelParams, PipelineParams, get_combined_args
from gaussian_renderer import GaussianModel
from utils.spiral_utils import spiral_cam_info
import re
import imageio
import numpy as np
from PIL import Image, ImageDraw
import copy
import random

def generate_video(imgs_path, text_to_add = ''):

    image_files = [f for f in os.listdir(imgs_path) if re.match(r'\d+\.png$', f) and '_depth' not in f]
    image_files_sorted = sorted(image_files, key=lambda x: int(re.findall(r'\d+', x)[0]))
    writer = imageio.get_writer(f"{imgs_path}/0000_video.mp4", fps=20)
    for img_name in image_files_sorted:
        img_path = os.path.join(imgs_path, img_name)
        img = Image.open(img_path)
        # Draw text on the image
        draw = ImageDraw.Draw(img)
        draw.text(( img.width - 10 * len(text_to_add), 20), text_to_add, fill="green" )  # Adjust position

        writer.append_data(np.array(img))
    writer.close()


def render_set(model_path, name, iteration, views, gaussians, pipeline, background, light, loaded_iter):
    main_path = os.path.join(model_path, name, "ours_{}".format(iteration))
    render_path = os.path.join(main_path, "renders")
    gts_path = os.path.join(main_path, "gt")
    seaprate_path_fixed_cam = os.path.join(main_path, "separate_light_camera_fixed_cam")
    seaprate_path_fixed_light = os.path.join(main_path, "separate_light_camera_fixed_light")
    disable_reflections_path = os.path.join(main_path, "disable_reflections")
    normal_path = os.path.join(main_path, "normals")
    albedo_path = os.path.join(main_path, "albedo")
    views = sorted(views, key=lambda obj: obj.colmap_id)

    makedirs(render_path, exist_ok=True)
    makedirs(gts_path, exist_ok=True)
    makedirs(disable_reflections_path, exist_ok=True)
    makedirs(albedo_path, exist_ok=True)
    makedirs(normal_path, exist_ok=True)

    middle_cam_idx = int(len(views)//2)
    middle_cam = views[middle_cam_idx]
    first_cam = views[0]

    random.seed(42)
    random_ids = random.sample(range(len(views)), min(20,len(views)))

    for idx, view in enumerate(tqdm(views, desc="Rendering progress")):

        gt = view.original_image.cuda()
        torchvision.utils.save_image(gt, os.path.join(gts_path, '{0:05d}'.format(idx) + ".png"))

        override_color = gaussians.compute_lighted_rgb(camera_center = view.camera_center, light = light, iter = loaded_iter)
        renderpkg = render(view, gaussians, pipeline, background, override_color = override_color)
        rendering = renderpkg["render"].clamp(0,1)
        torchvision.utils.save_image(rendering, os.path.join(render_path, '{0:05d}'.format(idx) + ".png"))

        if True: #idx in random_ids:
            override_color = gaussians.compute_lighted_rgb(camera_center = view.camera_center, light = light, iter = loaded_iter, disable_reflections=True)
            renderpkg = render(view, gaussians, pipeline, background, override_color = override_color)
            rendering = renderpkg["render"].clamp(0,1)
            torchvision.utils.save_image(rendering, os.path.join(disable_reflections_path, '{0:05d}'.format(idx) + ".png"))

            override_color = gaussians.compute_gaussian_rgb(view.camera_center)
            renderpkg = render(view, gaussians, pipeline, background, override_color = override_color)
            rendering = renderpkg["render"].clamp(min=1e-8, max=1)**(1/2.2) # linear to srgb
            torchvision.utils.save_image(rendering, os.path.join(albedo_path, '{0:05d}'.format(idx) + ".png"))

            # render normals
            dir_pp_camera = (gaussians.get_xyz - view.camera_center.repeat(gaussians.get_features.shape[0], 1))
            normal = gaussians.get_gaussian_normals()[:, 3:]

            # # Normalize the input vectors
            N = torch.nn.functional.normalize(normal, dim=1)
            V = torch.nn.functional.normalize(dir_pp_camera, dim=1)

            # # normals always towards camera
            N_dot_V = torch.sum(N * -V, dim=1, keepdim=True)  # [N, 1]
            N = torch.where(N_dot_V < 0, -N, N)  # Flip N if N_dot_V < 0
            override_color = (N+1)/2
            renderpkg = render(view, gaussians, pipeline, background, override_color = override_color)
            rendering = renderpkg["render"].clamp(0,1)
            torchvision.utils.save_image(rendering, os.path.join(normal_path, '{0:05d}'.format(idx) + ".png"))


    # Make videos
    generate_video(render_path, "renders")
    generate_video(albedo_path, "albedo")
    generate_video(disable_reflections_path, "disable_refl")
    generate_video(normal_path, "normals")


    # Performence check
    render_time_list = []
    for idx, view in enumerate(tqdm(views, desc="Rendering progress")):
        torch.cuda.synchronize()
        start = time.time()
        override_color = gaussians.compute_lighted_rgb(camera_center = view.camera_center, light = light, iter = loaded_iter)
        renderpkg = render(view, gaussians, pipeline, background, override_color = override_color)

        torch.cuda.synchronize()
        end = time.time()
        render_time_list.append((end-start)*1000)
    with open(os.path.join(model_path, 'render_time.txt'), 'w') as f:
        for t in render_time_list:
            f.write("%.2fms\n"%t)
        f.write("Mean time: %.2fms\n"%(np.mean(render_time_list[5:])))

def render_sets(dataset : ModelParams, iteration : int, pipeline : PipelineParams, skip_train : bool, skip_test : bool, skip_spiral : bool):
    with torch.no_grad():
        gaussians = GaussianModel(dataset.sh_degree)
        scene = Scene(dataset, gaussians, load_iteration=iteration, shuffle=False)

        bg_color = [1,1,1] if dataset.white_background else [0, 0, 0]
        background = torch.tensor(bg_color, dtype=torch.float32, device="cuda")
        gaussians.mlp.eval()
        gaussians.positional_encoding_camera.eval()

        if not skip_train:
             render_set(dataset.model_path, "train", scene.loaded_iter, scene.getTrainCameras(), gaussians, pipeline, background, scene.light, scene.loaded_iter)

        if not skip_test:
             render_set(dataset.model_path, "test", scene.loaded_iter, scene.getTestCameras(), gaussians, pipeline, background, scene.light, scene.loaded_iter)

        # if not skip_spiral:
        #     cams = spiral_cam_info(dataset.model_path, focal=0.7)
        #     render_set(dataset.model_path, "spiral", scene.loaded_iter,cams, gaussians, pipeline, background)

if __name__ == "__main__":
    # Set up command line argument parser
    parser = ArgumentParser(description="Testing script parameters")
    model = ModelParams(parser, sentinel=True)
    pipeline = PipelineParams(parser)

    parser.add_argument("--iteration", default=-1, type=int)
    parser.add_argument("--skip_train", action="store_true")
    parser.add_argument("--skip_test", action="store_true")
    parser.add_argument("--skip_spiral", action="store_true")
    parser.add_argument("--quiet", action="store_true")
    args = get_combined_args(parser)
    print("Rendering " + args.model_path)

    # Initialize system state (RNG)
    safe_state(args.quiet)


    render_sets(model.extract(args), args.iteration, pipeline.extract(args), args.skip_train, args.skip_test, args.skip_spiral)