-
Notifications
You must be signed in to change notification settings - Fork 1k
/
img2tensorboard.py
211 lines (182 loc) · 8.98 KB
/
img2tensorboard.py
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
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
# Copyright (c) MONAI Consortium
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# http://www.apache.org/licenses/LICENSE-2.0
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import annotations
from typing import TYPE_CHECKING, Any
import numpy as np
import torch
from monai.config import NdarrayTensor
from monai.transforms import rescale_array
from monai.utils import convert_data_type, optional_import
PIL, _ = optional_import("PIL")
GifImage, _ = optional_import("PIL.GifImagePlugin", name="Image")
if TYPE_CHECKING:
from tensorboard.compat.proto.summary_pb2 import Summary
from tensorboardX import SummaryWriter as SummaryWriterX
from tensorboardX.proto.summary_pb2 import Summary as SummaryX
from torch.utils.tensorboard import SummaryWriter
has_tensorboardx = True
else:
Summary, _ = optional_import("tensorboard.compat.proto.summary_pb2", name="Summary")
SummaryX, _ = optional_import("tensorboardX.proto.summary_pb2", name="Summary")
SummaryWriter, _ = optional_import("torch.utils.tensorboard", name="SummaryWriter")
SummaryWriterX, has_tensorboardx = optional_import("tensorboardX", name="SummaryWriter")
__all__ = ["make_animated_gif_summary", "add_animated_gif", "plot_2d_or_3d_image"]
def _image3_animated_gif(
tag: str,
image: np.ndarray | torch.Tensor,
writer: SummaryWriter | SummaryWriterX | None,
frame_dim: int = 0,
scale_factor: float = 1.0,
) -> Any:
"""Function to actually create the animated gif.
Args:
tag: Data identifier
image: 3D image tensors expected to be in `HWD` format
writer: the tensorboard writer to plot image
frame_dim: the dimension used as frames for GIF image, expect data shape as `HWD`, default to `0`.
scale_factor: amount to multiply values by. if the image data is between 0 and 1, using 255 for this value will
scale it to displayable range
"""
if len(image.shape) != 3:
raise AssertionError("3D image tensors expected to be in `HWD` format, len(image.shape) != 3")
image_np, *_ = convert_data_type(image, output_type=np.ndarray)
ims = [(i * scale_factor).astype(np.uint8, copy=False) for i in np.moveaxis(image_np, frame_dim, 0)]
ims = [GifImage.fromarray(im) for im in ims]
img_str = b""
for b_data in PIL.GifImagePlugin.getheader(ims[0])[0]:
img_str += b_data
img_str += b"\x21\xFF\x0B\x4E\x45\x54\x53\x43\x41\x50" b"\x45\x32\x2E\x30\x03\x01\x00\x00\x00"
for i in ims:
for b_data in PIL.GifImagePlugin.getdata(i):
img_str += b_data
img_str += b"\x3B"
summary = SummaryX if has_tensorboardx and isinstance(writer, SummaryWriterX) else Summary
summary_image_str = summary.Image(height=10, width=10, colorspace=1, encoded_image_string=img_str)
image_summary = summary.Value(tag=tag, image=summary_image_str)
return summary(value=[image_summary])
def make_animated_gif_summary(
tag: str,
image: np.ndarray | torch.Tensor,
writer: SummaryWriter | SummaryWriterX | None = None,
max_out: int = 3,
frame_dim: int = -3,
scale_factor: float = 1.0,
) -> Summary:
"""Creates an animated gif out of an image tensor in 'CHWD' format and returns Summary.
Args:
tag: Data identifier
image: The image, expected to be in `CHWD` format
writer: the tensorboard writer to plot image
max_out: maximum number of image channels to animate through
frame_dim: the dimension used as frames for GIF image, expect input data shape as `CHWD`,
default to `-3` (the first spatial dim)
scale_factor: amount to multiply values by.
if the image data is between 0 and 1, using 255 for this value will scale it to displayable range
"""
suffix = "/image" if max_out == 1 else "/image/{}"
# GIF image has no channel dim, reduce the spatial dim index if positive
frame_dim = frame_dim - 1 if frame_dim > 0 else frame_dim
summary_op = []
for it_i in range(min(max_out, list(image.shape)[0])):
one_channel_img: torch.Tensor | np.ndarray = (
image[it_i, :, :, :].squeeze(dim=0) if isinstance(image, torch.Tensor) else image[it_i, :, :, :]
)
summary_op.append(
_image3_animated_gif(tag + suffix.format(it_i), one_channel_img, writer, frame_dim, scale_factor)
)
return summary_op
def add_animated_gif(
writer: SummaryWriter | SummaryWriterX,
tag: str,
image_tensor: np.ndarray | torch.Tensor,
max_out: int = 3,
frame_dim: int = -3,
scale_factor: float = 1.0,
global_step: int | None = None,
) -> None:
"""Creates an animated gif out of an image tensor in 'CHWD' format and writes it with SummaryWriter.
Args:
writer: Tensorboard SummaryWriter to write to
tag: Data identifier
image_tensor: tensor for the image to add, expected to be in `CHWD` format
max_out: maximum number of image channels to animate through
frame_dim: the dimension used as frames for GIF image, expect input data shape as `CHWD`,
default to `-3` (the first spatial dim)
scale_factor: amount to multiply values by. If the image data is between 0 and 1, using 255 for this value will
scale it to displayable range
global_step: Global step value to record
"""
summary = make_animated_gif_summary(
tag=tag, image=image_tensor, writer=writer, max_out=max_out, frame_dim=frame_dim, scale_factor=scale_factor
)
for s in summary:
# add GIF for every channel separately
writer._get_file_writer().add_summary(s, global_step)
def plot_2d_or_3d_image(
data: NdarrayTensor | list[NdarrayTensor],
step: int,
writer: SummaryWriter | SummaryWriterX,
index: int = 0,
max_channels: int = 1,
frame_dim: int = -3,
max_frames: int = 24,
tag: str = "output",
) -> None:
"""Plot 2D or 3D image on the TensorBoard, 3D image will be converted to GIF image.
Note:
Plot 3D or 2D image(with more than 3 channels) as separate images.
And if writer is from TensorBoardX, data has 3 channels and `max_channels=3`, will plot as RGB video.
Args:
data: target data to be plotted as image on the TensorBoard.
The data is expected to have 'NCHW[D]' dimensions or a list of data with `CHW[D]` dimensions,
and only plot the first in the batch.
step: current step to plot in a chart.
writer: specify TensorBoard or TensorBoardX SummaryWriter to plot the image.
index: plot which element in the input data batch, default is the first element.
max_channels: number of channels to plot.
frame_dim: if plotting 3D image as GIF, specify the dimension used as frames,
expect input data shape as `NCHWD`, default to `-3` (the first spatial dim)
max_frames: if plot 3D RGB image as video in TensorBoardX, set the FPS to `max_frames`.
tag: tag of the plotted image on TensorBoard.
"""
data_index = data[index]
# as the `d` data has no batch dim, reduce the spatial dim index if positive
frame_dim = frame_dim - 1 if frame_dim > 0 else frame_dim
d: np.ndarray = data_index.detach().cpu().numpy() if isinstance(data_index, torch.Tensor) else data_index
if d.ndim == 2:
d = rescale_array(d, 0, 1) # type: ignore
dataformats = "HW"
writer.add_image(f"{tag}_{dataformats}", d, step, dataformats=dataformats)
return
if d.ndim == 3:
if d.shape[0] == 3 and max_channels == 3: # RGB
dataformats = "CHW"
writer.add_image(f"{tag}_{dataformats}", d, step, dataformats=dataformats)
return
dataformats = "HW"
for j, d2 in enumerate(d[:max_channels]):
d2 = rescale_array(d2, 0, 1)
writer.add_image(f"{tag}_{dataformats}_{j}", d2, step, dataformats=dataformats)
return
if d.ndim >= 4:
spatial = d.shape[-3:]
d = d.reshape([-1] + list(spatial))
if d.shape[0] == 3 and max_channels == 3 and has_tensorboardx and isinstance(writer, SummaryWriterX): # RGB
# move the expected frame dim to the end as `T` dim for video
d = np.moveaxis(d, frame_dim, -1)
writer.add_video(tag, d[None], step, fps=max_frames, dataformats="NCHWT")
return
# scale data to 0 - 255 for visualization
max_channels = min(max_channels, d.shape[0])
d = np.stack([rescale_array(i, 0, 255) for i in d[:max_channels]], axis=0)
# will plot every channel as a separate GIF image
add_animated_gif(writer, f"{tag}_HWD", d, max_out=max_channels, frame_dim=frame_dim, global_step=step)
return