File size: 5,591 Bytes
55ed985
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
import base64
import logging
import math
import os
import subprocess
from glob import glob
from io import BytesIO
from typing import Union

import cv2
import imageio
import numpy as np
import PIL.Image as Image
from moviepy.editor import VideoFileClip, clips_array

logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)


__all__ = [
    "render_asset3d",
    "merge_images_video",
    "filter_small_connected_components",
    "filter_image_small_connected_components",
    "combine_images_to_base64",
]


def render_asset3d(
    mesh_path: str,
    output_root: str,
    distance: float = 5.0,
    num_images: int = 1,
    elevation: list[float] = (0.0,),
    pbr_light_factor: float = 1.5,
    return_key: str = "image_color/*",
    output_subdir: str = "renders",
    gen_color_mp4: bool = False,
    gen_viewnormal_mp4: bool = False,
    gen_glonormal_mp4: bool = False,
    device: str = "cpu",
) -> list[str]:
    command = [
        "python3",
        "asset3d_gen/data/differentiable_render.py",
        "--mesh_path",
        mesh_path,
        "--output_root",
        output_root,
        "--uuid",
        output_subdir,
        "--distance",
        str(distance),
        "--num_images",
        str(num_images),
        "--elevation",
        *map(str, elevation),
        "--pbr_light_factor",
        str(pbr_light_factor),
        "--with_mtl",
        "--device",
        device,
    ]
    if gen_color_mp4:
        command.append("--gen_color_mp4")
    if gen_viewnormal_mp4:
        command.append("--gen_viewnormal_mp4")
    if gen_glonormal_mp4:
        command.append("--gen_glonormal_mp4")
    try:
        subprocess.run(command, check=True)
    except subprocess.CalledProcessError as e:
        logger.error(f"Error occurred during rendering: {e}.")

    dst_paths = glob(os.path.join(output_root, output_subdir, return_key))

    return dst_paths


def merge_images_video(color_images, normal_images, output_path) -> None:
    width = color_images[0].shape[1]
    combined_video = [
        np.hstack([rgb_img[:, : width // 2], normal_img[:, width // 2 :]])
        for rgb_img, normal_img in zip(color_images, normal_images)
    ]
    imageio.mimsave(output_path, combined_video, fps=50)

    return


def merge_video_video(
    video_path1: str, video_path2: str, output_path: str
) -> None:
    """Merge two videos by the left half and the right half of the videos."""
    clip1 = VideoFileClip(video_path1)
    clip2 = VideoFileClip(video_path2)

    if clip1.size != clip2.size:
        raise ValueError("The resolutions of the two videos do not match.")

    width, height = clip1.size
    clip1_half = clip1.crop(x1=0, y1=0, x2=width // 2, y2=height)
    clip2_half = clip2.crop(x1=width // 2, y1=0, x2=width, y2=height)
    final_clip = clips_array([[clip1_half, clip2_half]])
    final_clip.write_videofile(output_path, codec="libx264")


def filter_small_connected_components(
    mask: Union[Image.Image, np.ndarray],
    area_ratio: float,
    connectivity: int = 8,
) -> np.ndarray:
    if isinstance(mask, Image.Image):
        mask = np.array(mask)
    num_labels, labels, stats, _ = cv2.connectedComponentsWithStats(
        mask,
        connectivity=connectivity,
    )

    small_components = np.zeros_like(mask, dtype=np.uint8)
    mask_area = (mask != 0).sum()
    min_area = mask_area // area_ratio
    for label in range(1, num_labels):
        area = stats[label, cv2.CC_STAT_AREA]
        if area < min_area:
            small_components[labels == label] = 255

    mask = cv2.bitwise_and(mask, cv2.bitwise_not(small_components))

    return mask


def filter_image_small_connected_components(
    image: Union[Image.Image, np.ndarray],
    area_ratio: float = 10,
    connectivity: int = 8,
) -> np.ndarray:
    if isinstance(image, Image.Image):
        image = image.convert("RGBA")
        image = np.array(image)

    mask = image[..., 3]
    mask = filter_small_connected_components(mask, area_ratio, connectivity)
    image[..., 3] = mask

    return image


def combine_images_to_base64(
    images: list[str | Image.Image],
    cat_row_col: tuple[int, int] = None,
    target_wh: tuple[int, int] = (512, 512),
) -> str:
    n_images = len(images)
    if cat_row_col is None:
        n_col = math.ceil(math.sqrt(n_images))
        n_row = math.ceil(n_images / n_col)
    else:
        n_row, n_col = cat_row_col

    images = [
        Image.open(p).convert("RGB") if isinstance(p, str) else p
        for p in images[: n_row * n_col]
    ]
    images = [img.resize(target_wh) for img in images]

    grid_w, grid_h = n_col * target_wh[0], n_row * target_wh[1]
    grid = Image.new("RGB", (grid_w, grid_h), (255, 255, 255))

    for idx, img in enumerate(images):
        row, col = divmod(idx, n_col)
        grid.paste(img, (col * target_wh[0], row * target_wh[1]))

    buffer = BytesIO()
    grid.save(buffer, format="PNG")

    return base64.b64encode(buffer.getvalue()).decode("utf-8")


if __name__ == "__main__":
    # Example usage:
    merge_video_video(
        "outputs/imageto3d/room_bottle7/room_bottle_007/URDF_room_bottle_007/mesh_glo_normal.mp4",  # noqa
        "outputs/imageto3d/room_bottle7/room_bottle_007/URDF_room_bottle_007/mesh.mp4",  # noqa
        "merge.mp4",
    )

    image_base64 = combine_images_to_base64(
        [
            "outputs/text2image/demo_objects/bed/sample_0.jpg",
            "outputs/imageto3d/v2/cups/sample_69/URDF_sample_69/qa_renders/image_color/003.png",  # noqa
            "outputs/text2image/demo_objects/cardboard/sample_1.jpg",
        ]
    )