Commit
·
eadd412
1
Parent(s):
0c7f2c0
fix eval and plotting refactor
Browse files- .gitignore +2 -0
- README.md +1 -1
- eval.py +38 -60
- fid_score.py +3 -2
- main.py +3 -187
- plots.py +254 -0
.gitignore
CHANGED
@@ -3,3 +3,5 @@
|
|
3 |
temp/
|
4 |
*.png
|
5 |
*.pdf
|
|
|
|
|
|
3 |
temp/
|
4 |
*.png
|
5 |
*.pdf
|
6 |
+
*.hash
|
7 |
+
*.npz
|
README.md
CHANGED
@@ -7,7 +7,7 @@
|
|
7 |
<a href="https://keras.io/"><img src="https://img.shields.io/badge/Keras-EE4C2C?logo=keras&logoColor=white" alt="Keras"></a>
|
8 |
</p>
|
9 |
<h3>
|
10 |
-
<span style="display:inline-block; margin: 0
|
11 |
<a href="https://example.com/tristan-stevens">Tristan Stevens</a>
|
12 |
</span>
|
13 |
<span style="display:inline-block; margin: 0 20px;">
|
|
|
7 |
<a href="https://keras.io/"><img src="https://img.shields.io/badge/Keras-EE4C2C?logo=keras&logoColor=white" alt="Keras"></a>
|
8 |
</p>
|
9 |
<h3>
|
10 |
+
<span style="display:inline-block; margin: 0 40px;">
|
11 |
<a href="https://example.com/tristan-stevens">Tristan Stevens</a>
|
12 |
</span>
|
13 |
<span style="display:inline-block; margin: 0 20px;">
|
eval.py
CHANGED
@@ -2,16 +2,17 @@ import warnings
|
|
2 |
from glob import glob
|
3 |
from pathlib import Path
|
4 |
|
5 |
-
import matplotlib.pyplot as plt
|
6 |
import numpy as np
|
7 |
import torch
|
8 |
import tyro
|
9 |
from PIL import Image
|
10 |
from scipy.ndimage import binary_erosion, distance_transform_edt
|
11 |
from scipy.stats import ks_2samp
|
|
|
12 |
from zea.io_lib import load_image
|
13 |
|
14 |
import fid_score
|
|
|
15 |
|
16 |
|
17 |
def calculate_fid_score(denoised_image_dirs, ground_truth_dir):
|
@@ -207,64 +208,38 @@ def calculate_final_score(aggregates):
|
|
207 |
return 0
|
208 |
|
209 |
|
210 |
-
def
|
211 |
-
|
212 |
-
|
213 |
-
|
214 |
-
|
215 |
-
|
216 |
-
|
217 |
-
|
218 |
-
|
219 |
-
|
220 |
-
|
221 |
-
|
222 |
-
|
223 |
-
values,
|
224 |
-
bins=30,
|
225 |
-
color=colors[idx % len(colors)],
|
226 |
-
alpha=0.85,
|
227 |
-
edgecolor="black",
|
228 |
-
linewidth=0.7,
|
229 |
-
)
|
230 |
-
ax.set_xlabel(metric_labels.get(name, name), fontsize=11)
|
231 |
-
ax.set_ylabel("Count", fontsize=10)
|
232 |
-
# Draw limits
|
233 |
-
if name in limits:
|
234 |
-
for lim in limits[name]:
|
235 |
-
ax.axvline(lim, color="crimson", linestyle="--", lw=1.2)
|
236 |
-
ax.spines["top"].set_visible(False)
|
237 |
-
ax.spines["right"].set_visible(False)
|
238 |
-
ax.tick_params(axis="both", which="major", labelsize=9)
|
239 |
-
fig.tight_layout(pad=1.5)
|
240 |
-
fig.savefig(out_path, bbox_inches="tight", dpi=600)
|
241 |
-
plt.close(fig)
|
242 |
-
|
243 |
-
|
244 |
-
def main(folder: str, roi_folder: str, reference_folder: str):
|
245 |
folder = Path(folder)
|
|
|
246 |
roi_folder = Path(roi_folder)
|
247 |
reference_folder = Path(reference_folder)
|
248 |
|
249 |
folder_files = set(f.name for f in folder.glob("*.png"))
|
|
|
250 |
roi_files = set(f.name for f in roi_folder.glob("*.png"))
|
251 |
-
ref_files = set(f.name for f in reference_folder.glob("*.png"))
|
252 |
|
253 |
print(f"Found {len(folder_files)} .png files in output folder: {folder}")
|
|
|
254 |
print(f"Found {len(roi_files)} .png files in ROI folder: {roi_folder}")
|
255 |
-
print(f"Found {len(ref_files)} .png files in reference folder: {reference_folder}")
|
256 |
|
257 |
# Find intersection of filenames
|
258 |
-
common_files = sorted(folder_files & roi_files &
|
259 |
-
print(f"Found {len(common_files)} images
|
260 |
-
|
261 |
-
|
262 |
-
|
263 |
-
print(f"ROI folder files: {sorted(roi_files)}")
|
264 |
-
print(f"Reference folder files: {sorted(ref_files)}")
|
265 |
-
assert len(common_files) > 0, (
|
266 |
-
"No matching .png files in all folders. Cannot proceed."
|
267 |
-
)
|
268 |
|
269 |
metrics = {"CNR": [], "gCNR": [], "KS_A": [], "KS_B": []}
|
270 |
limits = {
|
@@ -275,28 +250,26 @@ def main(folder: str, roi_folder: str, reference_folder: str):
|
|
275 |
}
|
276 |
|
277 |
for name in common_files:
|
278 |
-
|
|
|
279 |
roi_path = roi_folder / name
|
280 |
-
ref_path = reference_folder / name
|
281 |
-
|
282 |
-
assert our_path.exists(), f"Missing file in output folder: {our_path}"
|
283 |
-
assert roi_path.exists(), f"Missing file in ROI folder: {roi_path}"
|
284 |
-
assert ref_path.exists(), f"Missing file in reference folder: {ref_path}"
|
285 |
|
286 |
try:
|
287 |
-
|
288 |
-
|
289 |
except Exception as e:
|
290 |
print(f"Error loading image {name}: {e}")
|
291 |
continue
|
292 |
|
293 |
# CNR/gCNR
|
294 |
-
cnr_gcnr = calculate_cnr_gcnr(
|
295 |
metrics["CNR"].append(cnr_gcnr[0][0])
|
296 |
metrics["gCNR"].append(cnr_gcnr[0][1])
|
297 |
|
298 |
# KS statistics
|
299 |
-
ks_a, _, ks_b, _ = calculate_ks_statistics(
|
|
|
|
|
300 |
metrics["KS_A"].append(ks_a)
|
301 |
metrics["KS_B"].append(ks_b)
|
302 |
|
@@ -308,8 +281,13 @@ def main(folder: str, roi_folder: str, reference_folder: str):
|
|
308 |
for k, (mean, std, minv, maxv) in stats.items():
|
309 |
print(f"{k}: mean={mean:.3f}, std={std:.3f}, min={minv:.3f}, max={maxv:.3f}")
|
310 |
|
311 |
-
plot_metrics(metrics, limits,
|
312 |
-
|
|
|
|
|
|
|
|
|
|
|
313 |
|
314 |
# Compute FID
|
315 |
fid_image_paths = [str(folder / name) for name in common_files]
|
|
|
2 |
from glob import glob
|
3 |
from pathlib import Path
|
4 |
|
|
|
5 |
import numpy as np
|
6 |
import torch
|
7 |
import tyro
|
8 |
from PIL import Image
|
9 |
from scipy.ndimage import binary_erosion, distance_transform_edt
|
10 |
from scipy.stats import ks_2samp
|
11 |
+
from zea import log
|
12 |
from zea.io_lib import load_image
|
13 |
|
14 |
import fid_score
|
15 |
+
from plots import plot_metrics
|
16 |
|
17 |
|
18 |
def calculate_fid_score(denoised_image_dirs, ground_truth_dir):
|
|
|
208 |
return 0
|
209 |
|
210 |
|
211 |
+
def main(folder: str, noisy_folder: str, roi_folder: str, reference_folder: str):
|
212 |
+
"""Evaluate the dehazing algorithm.
|
213 |
+
|
214 |
+
Args:
|
215 |
+
folder (str): Path to the folder containing the dehazed images.
|
216 |
+
Used for evaluating all metrics.
|
217 |
+
noisy_folder (str): Path to the folder containing the noisy images.
|
218 |
+
Only used for KS statistics.
|
219 |
+
roi_folder (str): Path to the folder containing the ROI images.
|
220 |
+
Used for contrast and KS statistic metrics.
|
221 |
+
reference_folder (str): Path to the folder containing the reference images.
|
222 |
+
Used only for FID calculation.
|
223 |
+
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
224 |
folder = Path(folder)
|
225 |
+
noisy_folder = Path(noisy_folder)
|
226 |
roi_folder = Path(roi_folder)
|
227 |
reference_folder = Path(reference_folder)
|
228 |
|
229 |
folder_files = set(f.name for f in folder.glob("*.png"))
|
230 |
+
noisy_files = set(f.name for f in noisy_folder.glob("*.png"))
|
231 |
roi_files = set(f.name for f in roi_folder.glob("*.png"))
|
|
|
232 |
|
233 |
print(f"Found {len(folder_files)} .png files in output folder: {folder}")
|
234 |
+
print(f"Found {len(noisy_files)} .png files in noisy folder: {noisy_folder}")
|
235 |
print(f"Found {len(roi_files)} .png files in ROI folder: {roi_folder}")
|
|
|
236 |
|
237 |
# Find intersection of filenames
|
238 |
+
common_files = sorted(folder_files & roi_files & noisy_files)
|
239 |
+
print(f"Found {len(common_files)} matching images in noisy/dehazed/roi folders")
|
240 |
+
assert len(common_files) > 0, (
|
241 |
+
"No matching .png files in all folders. Cannot proceed."
|
242 |
+
)
|
|
|
|
|
|
|
|
|
|
|
243 |
|
244 |
metrics = {"CNR": [], "gCNR": [], "KS_A": [], "KS_B": []}
|
245 |
limits = {
|
|
|
250 |
}
|
251 |
|
252 |
for name in common_files:
|
253 |
+
dehazed_path = folder / name
|
254 |
+
noisy_path = noisy_folder / name
|
255 |
roi_path = roi_folder / name
|
|
|
|
|
|
|
|
|
|
|
256 |
|
257 |
try:
|
258 |
+
img_dehazed = np.array(load_image(str(dehazed_path)))
|
259 |
+
img_noisy = np.array(load_image(str(noisy_path)))
|
260 |
except Exception as e:
|
261 |
print(f"Error loading image {name}: {e}")
|
262 |
continue
|
263 |
|
264 |
# CNR/gCNR
|
265 |
+
cnr_gcnr = calculate_cnr_gcnr(img_dehazed, str(roi_path))
|
266 |
metrics["CNR"].append(cnr_gcnr[0][0])
|
267 |
metrics["gCNR"].append(cnr_gcnr[0][1])
|
268 |
|
269 |
# KS statistics
|
270 |
+
ks_a, _, ks_b, _ = calculate_ks_statistics(
|
271 |
+
img_noisy, img_dehazed, str(roi_path)
|
272 |
+
)
|
273 |
metrics["KS_A"].append(ks_a)
|
274 |
metrics["KS_B"].append(ks_b)
|
275 |
|
|
|
281 |
for k, (mean, std, minv, maxv) in stats.items():
|
282 |
print(f"{k}: mean={mean:.3f}, std={std:.3f}, min={minv:.3f}, max={maxv:.3f}")
|
283 |
|
284 |
+
fig = plot_metrics(metrics, limits, "contrast_metrics.png")
|
285 |
+
|
286 |
+
path = Path("contrast_metrics.png")
|
287 |
+
save_kwargs = {"bbox_inches": "tight", "dpi": 300}
|
288 |
+
fig.savefig(path, **save_kwargs)
|
289 |
+
fig.savefig(path.with_suffix(".pdf"), **save_kwargs)
|
290 |
+
log.success(f"Metrics plot saved to {log.yellow(path)}")
|
291 |
|
292 |
# Compute FID
|
293 |
fid_image_paths = [str(folder / name) for name in common_files]
|
fid_score.py
CHANGED
@@ -88,6 +88,7 @@ parser.add_argument(
|
|
88 |
)
|
89 |
|
90 |
IMAGE_EXTENSIONS = {"bmp", "jpg", "jpeg", "pgm", "png", "ppm", "tif", "tiff", "webp"}
|
|
|
91 |
|
92 |
|
93 |
class ImagePathDataset(torch.utils.data.Dataset):
|
@@ -277,7 +278,7 @@ def compute_statistics_of_path(path, model, batch_size, dims, device, num_worker
|
|
277 |
|
278 |
|
279 |
def _fid_cache_paths():
|
280 |
-
tmp_dir =
|
281 |
tmp_dir.mkdir(exist_ok=True)
|
282 |
stats_path = tmp_dir / "fid_stats.npz"
|
283 |
hash_path = tmp_dir / "fid_stats.hash"
|
@@ -391,7 +392,7 @@ def calculate_fid_with_cached_ground_truth(
|
|
391 |
continue
|
392 |
return hash_md5.hexdigest()
|
393 |
|
394 |
-
tmp_dir =
|
395 |
tmp_dir.mkdir(exist_ok=True)
|
396 |
stats_path = tmp_dir / "fid_stats.npz"
|
397 |
hash_path = tmp_dir / "fid_stats.hash"
|
|
|
88 |
)
|
89 |
|
90 |
IMAGE_EXTENSIONS = {"bmp", "jpg", "jpeg", "pgm", "png", "ppm", "tif", "tiff", "webp"}
|
91 |
+
TEMP_DIR = pathlib.Path("temp")
|
92 |
|
93 |
|
94 |
class ImagePathDataset(torch.utils.data.Dataset):
|
|
|
278 |
|
279 |
|
280 |
def _fid_cache_paths():
|
281 |
+
tmp_dir = TEMP_DIR
|
282 |
tmp_dir.mkdir(exist_ok=True)
|
283 |
stats_path = tmp_dir / "fid_stats.npz"
|
284 |
hash_path = tmp_dir / "fid_stats.hash"
|
|
|
392 |
continue
|
393 |
return hash_md5.hexdigest()
|
394 |
|
395 |
+
tmp_dir = TEMP_DIR
|
396 |
tmp_dir.mkdir(exist_ok=True)
|
397 |
stats_path = tmp_dir / "fid_stats.npz"
|
398 |
hash_path = tmp_dir / "fid_stats.hash"
|
main.py
CHANGED
@@ -1,9 +1,6 @@
|
|
1 |
import copy
|
2 |
-
import os
|
3 |
from pathlib import Path
|
4 |
|
5 |
-
os.environ["KERAS_BACKEND"] = "jax"
|
6 |
-
|
7 |
import jax
|
8 |
import keras
|
9 |
import matplotlib.pyplot as plt
|
@@ -12,10 +9,8 @@ import scipy
|
|
12 |
import tyro
|
13 |
import zea
|
14 |
from keras import ops
|
15 |
-
from matplotlib.patches import PathPatch
|
16 |
-
from matplotlib.path import Path as pltPath
|
17 |
from PIL import Image
|
18 |
-
from skimage import filters,
|
19 |
from zea import Config, init_device, log
|
20 |
from zea.internal.operators import Operator
|
21 |
from zea.models.diffusion import (
|
@@ -25,7 +20,8 @@ from zea.models.diffusion import (
|
|
25 |
)
|
26 |
from zea.tensor_ops import L2
|
27 |
from zea.utils import translate
|
28 |
-
|
|
|
29 |
|
30 |
|
31 |
def L1(x):
|
@@ -476,186 +472,6 @@ def run(
|
|
476 |
return hazy_images, pred_tissue_images, pred_haze_images, masks_out
|
477 |
|
478 |
|
479 |
-
def add_shape_from_mask(ax, mask, **kwargs):
|
480 |
-
"""add a shape to axis from mask array.
|
481 |
-
|
482 |
-
Args:
|
483 |
-
ax (plt.ax): matplotlib axis
|
484 |
-
mask (ndarray): numpy array with non-zero
|
485 |
-
shape defining the region of interest.
|
486 |
-
Kwargs:
|
487 |
-
edgecolor (str): color of the shape's edge
|
488 |
-
facecolor (str): color of the shape's face
|
489 |
-
linewidth (int): width of the shape's edge
|
490 |
-
|
491 |
-
Returns:
|
492 |
-
plt.ax: matplotlib axis with shape added
|
493 |
-
"""
|
494 |
-
# Pad mask to ensure edge contours are found
|
495 |
-
padded_mask = np.pad(mask, pad_width=1, mode="constant", constant_values=0)
|
496 |
-
contours = measure.find_contours(padded_mask, 0.5)
|
497 |
-
patches = []
|
498 |
-
for contour in contours:
|
499 |
-
# Remove padding offset
|
500 |
-
contour -= 1
|
501 |
-
path = pltPath(contour[:, ::-1])
|
502 |
-
patch = PathPatch(path, **kwargs)
|
503 |
-
patches.append(ax.add_patch(patch))
|
504 |
-
return patches
|
505 |
-
|
506 |
-
|
507 |
-
def plot_batch_with_named_masks(
|
508 |
-
images, masks_dict, mask_colors=None, titles=None, **kwargs
|
509 |
-
):
|
510 |
-
"""
|
511 |
-
Plot batch of images in rows, each column overlays a different mask from the dict.
|
512 |
-
Mask labels are shown as column titles. If mask name is 'per_pixel_omega', show it
|
513 |
-
directly with inferno colormap (no overlay).
|
514 |
-
|
515 |
-
Args:
|
516 |
-
images: np.ndarray, shape (batch, height, width, channels)
|
517 |
-
masks_dict: dict of {name: mask}, each mask shape (batch, height, width, channels)
|
518 |
-
mask_colors: dict of {name: color} or None (default colors used)
|
519 |
-
"""
|
520 |
-
mask_names = list(masks_dict.keys())
|
521 |
-
batch_size = images.shape[0]
|
522 |
-
default_colors = ["red", "green", "#33aaff", "yellow", "magenta", "cyan"]
|
523 |
-
mask_colors = mask_colors or {
|
524 |
-
name: default_colors[i % len(default_colors)]
|
525 |
-
for i, name in enumerate(mask_names)
|
526 |
-
}
|
527 |
-
|
528 |
-
# Prepare images for each column
|
529 |
-
columns = []
|
530 |
-
cmaps = []
|
531 |
-
for name in mask_names:
|
532 |
-
if name == "per_pixel_omega":
|
533 |
-
mask_np = np.array(masks_dict[name])
|
534 |
-
columns.append(np.squeeze(mask_np))
|
535 |
-
cmaps.append(["inferno"] * batch_size)
|
536 |
-
else:
|
537 |
-
columns.append(np.squeeze(images))
|
538 |
-
cmaps.append(["gray"] * batch_size)
|
539 |
-
|
540 |
-
# Stack columns: shape (num_columns, batch, ...)
|
541 |
-
all_images = np.stack(columns, axis=0) # (num_columns, batch, ...)
|
542 |
-
# Rearrange to (batch, num_columns, ...)
|
543 |
-
all_images = (
|
544 |
-
np.transpose(all_images, (1, 0, 2, 3, 4))
|
545 |
-
if all_images.ndim == 5
|
546 |
-
else np.transpose(all_images, (1, 0, 2, 3))
|
547 |
-
)
|
548 |
-
# Flatten to (batch * num_columns, ...)
|
549 |
-
all_images = all_images.reshape(batch_size * len(mask_names), *images.shape[1:])
|
550 |
-
|
551 |
-
# Flatten cmaps for plot_image_grid in the same order as images
|
552 |
-
flat_cmaps = []
|
553 |
-
for row in range(batch_size):
|
554 |
-
for col in range(len(mask_names)):
|
555 |
-
flat_cmaps.append(cmaps[col][row])
|
556 |
-
|
557 |
-
fig, _ = plot_image_grid(
|
558 |
-
all_images,
|
559 |
-
ncols=len(mask_names),
|
560 |
-
remove_axis=False,
|
561 |
-
cmap=flat_cmaps,
|
562 |
-
figsize=(8, 3.3),
|
563 |
-
**kwargs,
|
564 |
-
)
|
565 |
-
|
566 |
-
# Overlay masks for non-per_pixel_omega columns
|
567 |
-
for col_idx, name in enumerate(mask_names):
|
568 |
-
if name == "per_pixel_omega":
|
569 |
-
continue
|
570 |
-
mask_np = np.array(masks_dict[name])
|
571 |
-
axes = fig.axes[col_idx : batch_size * len(mask_names) : len(mask_names)]
|
572 |
-
for ax, mask_img in zip(axes, mask_np):
|
573 |
-
add_shape_from_mask(
|
574 |
-
ax, mask_img.squeeze(), color=mask_colors[name], alpha=0.3
|
575 |
-
)
|
576 |
-
|
577 |
-
# Add column titles
|
578 |
-
row_idx = 0
|
579 |
-
if titles is None:
|
580 |
-
titles = mask_names
|
581 |
-
for col_idx, name in enumerate(titles):
|
582 |
-
ax_idx = row_idx * len(mask_names) + col_idx
|
583 |
-
fig.axes[ax_idx].set_title(name, fontsize=9, color="white")
|
584 |
-
fig.axes[ax_idx].set_facecolor("black")
|
585 |
-
|
586 |
-
# Add colorbar for per_pixel_omega if present
|
587 |
-
if "per_pixel_omega" in mask_names:
|
588 |
-
col_idx = mask_names.index("per_pixel_omega")
|
589 |
-
axes = fig.axes[col_idx : batch_size * len(mask_names) : len(mask_names)]
|
590 |
-
|
591 |
-
# Get vertical bounds of the subplot column
|
592 |
-
top_ax = axes[0]
|
593 |
-
bottom_ax = axes[-1]
|
594 |
-
top_pos = top_ax.get_position()
|
595 |
-
bottom_pos = bottom_ax.get_position()
|
596 |
-
|
597 |
-
full_y0 = bottom_pos.y0
|
598 |
-
full_y1 = top_pos.y1
|
599 |
-
full_height = full_y1 - full_y0
|
600 |
-
|
601 |
-
# Manually shrink to 80% of full height and center vertically
|
602 |
-
scale = 0.8
|
603 |
-
height = full_height * scale
|
604 |
-
y0 = full_y0 + (full_height - height) / 2
|
605 |
-
|
606 |
-
x0 = top_pos.x1 + 0.015 # Horizontal position to the right
|
607 |
-
width = 0.015 # Thin bar
|
608 |
-
|
609 |
-
# Add colorbar axis
|
610 |
-
cax = fig.add_axes([x0, y0, width, height])
|
611 |
-
|
612 |
-
im = axes[0].get_images()[0] if axes[0].get_images() else None
|
613 |
-
cbar = fig.colorbar(im, cax=cax)
|
614 |
-
cbar.set_label(r"Guidance weighting \mathbf{p}")
|
615 |
-
cbar.ax.yaxis.set_major_locator(plt.MaxNLocator(nbins=6))
|
616 |
-
cbar.ax.yaxis.set_tick_params(labelsize=7)
|
617 |
-
cbar.ax.yaxis.label.set_size(8)
|
618 |
-
|
619 |
-
return fig
|
620 |
-
|
621 |
-
|
622 |
-
def plot_dehazed_results(
|
623 |
-
hazy_images,
|
624 |
-
pred_tissue_images,
|
625 |
-
pred_haze_images,
|
626 |
-
diffusion_model,
|
627 |
-
titles=("Hazy", "Dehazed", "Haze"),
|
628 |
-
):
|
629 |
-
"""Create and save visualization with optional mask overlays."""
|
630 |
-
|
631 |
-
# Create the processed image stack using the helper function
|
632 |
-
input_shape = diffusion_model.input_shape
|
633 |
-
stack_images = ops.stack(
|
634 |
-
[
|
635 |
-
hazy_images,
|
636 |
-
pred_tissue_images,
|
637 |
-
pred_haze_images,
|
638 |
-
]
|
639 |
-
)
|
640 |
-
stack_images = ops.reshape(stack_images, (-1, input_shape[0], input_shape[1]))
|
641 |
-
|
642 |
-
# Define labels based on what we're showing
|
643 |
-
fig, _ = plot_image_grid(
|
644 |
-
stack_images,
|
645 |
-
ncols=len(hazy_images),
|
646 |
-
remove_axis=False,
|
647 |
-
vmin=0,
|
648 |
-
vmax=255,
|
649 |
-
)
|
650 |
-
# Set labels and styling
|
651 |
-
for i, ax in enumerate(fig.axes):
|
652 |
-
if i % len(hazy_images) == 0:
|
653 |
-
label = titles[(i // len(hazy_images)) % len(titles)]
|
654 |
-
ax.set_ylabel(label, fontsize=12)
|
655 |
-
|
656 |
-
return fig
|
657 |
-
|
658 |
-
|
659 |
def main(
|
660 |
input_folder: str = "./assets",
|
661 |
output_folder: str = "./temp",
|
|
|
1 |
import copy
|
|
|
2 |
from pathlib import Path
|
3 |
|
|
|
|
|
4 |
import jax
|
5 |
import keras
|
6 |
import matplotlib.pyplot as plt
|
|
|
9 |
import tyro
|
10 |
import zea
|
11 |
from keras import ops
|
|
|
|
|
12 |
from PIL import Image
|
13 |
+
from skimage import filters, morphology
|
14 |
from zea import Config, init_device, log
|
15 |
from zea.internal.operators import Operator
|
16 |
from zea.models.diffusion import (
|
|
|
20 |
)
|
21 |
from zea.tensor_ops import L2
|
22 |
from zea.utils import translate
|
23 |
+
|
24 |
+
from plots import plot_batch_with_named_masks, plot_dehazed_results
|
25 |
|
26 |
|
27 |
def L1(x):
|
|
|
472 |
return hazy_images, pred_tissue_images, pred_haze_images, masks_out
|
473 |
|
474 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
475 |
def main(
|
476 |
input_folder: str = "./assets",
|
477 |
output_folder: str = "./temp",
|
plots.py
ADDED
@@ -0,0 +1,254 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import matplotlib.pyplot as plt
|
2 |
+
import numpy as np
|
3 |
+
from keras import ops
|
4 |
+
from matplotlib.patches import PathPatch
|
5 |
+
from matplotlib.path import Path as pltPath
|
6 |
+
from skimage import measure
|
7 |
+
from zea.visualize import plot_image_grid
|
8 |
+
|
9 |
+
|
10 |
+
def add_shape_from_mask(ax, mask, **kwargs):
|
11 |
+
"""add a shape to axis from mask array.
|
12 |
+
|
13 |
+
Args:
|
14 |
+
ax (plt.ax): matplotlib axis
|
15 |
+
mask (ndarray): numpy array with non-zero
|
16 |
+
shape defining the region of interest.
|
17 |
+
Kwargs:
|
18 |
+
edgecolor (str): color of the shape's edge
|
19 |
+
facecolor (str): color of the shape's face
|
20 |
+
linewidth (int): width of the shape's edge
|
21 |
+
|
22 |
+
Returns:
|
23 |
+
plt.ax: matplotlib axis with shape added
|
24 |
+
"""
|
25 |
+
# Pad mask to ensure edge contours are found
|
26 |
+
padded_mask = np.pad(mask, pad_width=1, mode="constant", constant_values=0)
|
27 |
+
contours = measure.find_contours(padded_mask, 0.5)
|
28 |
+
patches = []
|
29 |
+
for contour in contours:
|
30 |
+
# Remove padding offset
|
31 |
+
contour -= 1
|
32 |
+
path = pltPath(contour[:, ::-1])
|
33 |
+
patch = PathPatch(path, **kwargs)
|
34 |
+
patches.append(ax.add_patch(patch))
|
35 |
+
return patches
|
36 |
+
|
37 |
+
|
38 |
+
def plot_batch_with_named_masks(
|
39 |
+
images, masks_dict, mask_colors=None, titles=None, **kwargs
|
40 |
+
):
|
41 |
+
"""
|
42 |
+
Plot batch of images in rows, each column overlays a different mask from the dict.
|
43 |
+
Mask labels are shown as column titles. If mask name is 'per_pixel_omega', show it
|
44 |
+
directly with inferno colormap (no overlay).
|
45 |
+
|
46 |
+
Args:
|
47 |
+
images: np.ndarray, shape (batch, height, width, channels)
|
48 |
+
masks_dict: dict of {name: mask}, each mask shape (batch, height, width, channels)
|
49 |
+
mask_colors: dict of {name: color} or None (default colors used)
|
50 |
+
"""
|
51 |
+
mask_names = list(masks_dict.keys())
|
52 |
+
batch_size = images.shape[0]
|
53 |
+
default_colors = ["red", "green", "#33aaff", "yellow", "magenta", "cyan"]
|
54 |
+
mask_colors = mask_colors or {
|
55 |
+
name: default_colors[i % len(default_colors)]
|
56 |
+
for i, name in enumerate(mask_names)
|
57 |
+
}
|
58 |
+
|
59 |
+
# Prepare images for each column
|
60 |
+
columns = []
|
61 |
+
cmaps = []
|
62 |
+
for name in mask_names:
|
63 |
+
if name == "per_pixel_omega":
|
64 |
+
mask_np = np.array(masks_dict[name])
|
65 |
+
columns.append(np.squeeze(mask_np))
|
66 |
+
cmaps.append(["inferno"] * batch_size)
|
67 |
+
else:
|
68 |
+
columns.append(np.squeeze(images))
|
69 |
+
cmaps.append(["gray"] * batch_size)
|
70 |
+
|
71 |
+
# Stack columns: shape (num_columns, batch, ...)
|
72 |
+
all_images = np.stack(columns, axis=0) # (num_columns, batch, ...)
|
73 |
+
# Rearrange to (batch, num_columns, ...)
|
74 |
+
all_images = (
|
75 |
+
np.transpose(all_images, (1, 0, 2, 3, 4))
|
76 |
+
if all_images.ndim == 5
|
77 |
+
else np.transpose(all_images, (1, 0, 2, 3))
|
78 |
+
)
|
79 |
+
# Flatten to (batch * num_columns, ...)
|
80 |
+
all_images = all_images.reshape(batch_size * len(mask_names), *images.shape[1:])
|
81 |
+
|
82 |
+
# Flatten cmaps for plot_image_grid in the same order as images
|
83 |
+
flat_cmaps = []
|
84 |
+
for row in range(batch_size):
|
85 |
+
for col in range(len(mask_names)):
|
86 |
+
flat_cmaps.append(cmaps[col][row])
|
87 |
+
|
88 |
+
fig, _ = plot_image_grid(
|
89 |
+
all_images,
|
90 |
+
ncols=len(mask_names),
|
91 |
+
remove_axis=False,
|
92 |
+
cmap=flat_cmaps,
|
93 |
+
figsize=(8, 3.3),
|
94 |
+
**kwargs,
|
95 |
+
)
|
96 |
+
|
97 |
+
# Overlay masks for non-per_pixel_omega columns
|
98 |
+
for col_idx, name in enumerate(mask_names):
|
99 |
+
if name == "per_pixel_omega":
|
100 |
+
continue
|
101 |
+
mask_np = np.array(masks_dict[name])
|
102 |
+
axes = fig.axes[col_idx : batch_size * len(mask_names) : len(mask_names)]
|
103 |
+
for ax, mask_img in zip(axes, mask_np):
|
104 |
+
add_shape_from_mask(
|
105 |
+
ax, mask_img.squeeze(), color=mask_colors[name], alpha=0.3
|
106 |
+
)
|
107 |
+
|
108 |
+
# Add column titles
|
109 |
+
row_idx = 0
|
110 |
+
if titles is None:
|
111 |
+
titles = mask_names
|
112 |
+
for col_idx, name in enumerate(titles):
|
113 |
+
ax_idx = row_idx * len(mask_names) + col_idx
|
114 |
+
fig.axes[ax_idx].set_title(name, fontsize=9, color="white")
|
115 |
+
fig.axes[ax_idx].set_facecolor("black")
|
116 |
+
|
117 |
+
# Add colorbar for per_pixel_omega if present
|
118 |
+
if "per_pixel_omega" in mask_names:
|
119 |
+
col_idx = mask_names.index("per_pixel_omega")
|
120 |
+
axes = fig.axes[col_idx : batch_size * len(mask_names) : len(mask_names)]
|
121 |
+
|
122 |
+
# Get vertical bounds of the subplot column
|
123 |
+
top_ax = axes[0]
|
124 |
+
bottom_ax = axes[-1]
|
125 |
+
top_pos = top_ax.get_position()
|
126 |
+
bottom_pos = bottom_ax.get_position()
|
127 |
+
|
128 |
+
full_y0 = bottom_pos.y0
|
129 |
+
full_y1 = top_pos.y1
|
130 |
+
full_height = full_y1 - full_y0
|
131 |
+
|
132 |
+
# Manually shrink to 80% of full height and center vertically
|
133 |
+
scale = 0.8
|
134 |
+
height = full_height * scale
|
135 |
+
y0 = full_y0 + (full_height - height) / 2
|
136 |
+
|
137 |
+
x0 = top_pos.x1 + 0.015 # Horizontal position to the right
|
138 |
+
width = 0.015 # Thin bar
|
139 |
+
|
140 |
+
# Add colorbar axis
|
141 |
+
cax = fig.add_axes([x0, y0, width, height])
|
142 |
+
|
143 |
+
im = axes[0].get_images()[0] if axes[0].get_images() else None
|
144 |
+
cbar = fig.colorbar(im, cax=cax)
|
145 |
+
cbar.set_label(r"Guidance weighting \mathbf{p}")
|
146 |
+
cbar.ax.yaxis.set_major_locator(plt.MaxNLocator(nbins=6))
|
147 |
+
cbar.ax.yaxis.set_tick_params(labelsize=7)
|
148 |
+
cbar.ax.yaxis.label.set_size(8)
|
149 |
+
|
150 |
+
return fig
|
151 |
+
|
152 |
+
|
153 |
+
def plot_dehazed_results(
|
154 |
+
hazy_images,
|
155 |
+
pred_tissue_images,
|
156 |
+
pred_haze_images,
|
157 |
+
diffusion_model,
|
158 |
+
titles=("Hazy", "Dehazed", "Haze"),
|
159 |
+
):
|
160 |
+
"""Create and save visualization with optional mask overlays."""
|
161 |
+
|
162 |
+
# Create the processed image stack using the helper function
|
163 |
+
input_shape = diffusion_model.input_shape
|
164 |
+
stack_images = ops.stack(
|
165 |
+
[
|
166 |
+
hazy_images,
|
167 |
+
pred_tissue_images,
|
168 |
+
pred_haze_images,
|
169 |
+
]
|
170 |
+
)
|
171 |
+
stack_images = ops.reshape(stack_images, (-1, input_shape[0], input_shape[1]))
|
172 |
+
|
173 |
+
# Define labels based on what we're showing
|
174 |
+
fig, _ = plot_image_grid(
|
175 |
+
stack_images,
|
176 |
+
ncols=len(hazy_images),
|
177 |
+
remove_axis=False,
|
178 |
+
vmin=0,
|
179 |
+
vmax=255,
|
180 |
+
)
|
181 |
+
# Set labels and styling
|
182 |
+
for i, ax in enumerate(fig.axes):
|
183 |
+
if i % len(hazy_images) == 0:
|
184 |
+
label = titles[(i // len(hazy_images)) % len(titles)]
|
185 |
+
ax.set_ylabel(label, fontsize=12)
|
186 |
+
|
187 |
+
return fig
|
188 |
+
|
189 |
+
|
190 |
+
def plot_metrics(metrics, limits, out_path):
|
191 |
+
plt.style.use("seaborn-v0_8-darkgrid")
|
192 |
+
fig, axes = plt.subplots(1, len(metrics), figsize=(7.2, 2.7), dpi=600)
|
193 |
+
colors = ["#0057b7", "#ffb300", "#008744", "#d62d20"]
|
194 |
+
metric_labels = {
|
195 |
+
"CNR": r"CNR $\uparrow$",
|
196 |
+
"gCNR": r"gCNR $\uparrow$",
|
197 |
+
"KS_A": r"KS$_{septum}$ $\downarrow$",
|
198 |
+
"KS_B": r"KS$_{ventricle}$ $\uparrow$",
|
199 |
+
}
|
200 |
+
# For legend handles
|
201 |
+
legend_handles = []
|
202 |
+
import matplotlib.lines as mlines
|
203 |
+
|
204 |
+
min_style = {
|
205 |
+
"color": "crimson",
|
206 |
+
"linestyle": "--",
|
207 |
+
"lw": 1.2,
|
208 |
+
"marker": "o",
|
209 |
+
"markersize": 5,
|
210 |
+
}
|
211 |
+
max_style = {
|
212 |
+
"color": "crimson",
|
213 |
+
"linestyle": ":",
|
214 |
+
"lw": 1.2,
|
215 |
+
"marker": "s",
|
216 |
+
"markersize": 5,
|
217 |
+
}
|
218 |
+
for idx, (ax, (name, values)) in enumerate(zip(axes, metrics.items())):
|
219 |
+
ax.hist(
|
220 |
+
values,
|
221 |
+
bins=30,
|
222 |
+
color=colors[idx % len(colors)],
|
223 |
+
alpha=0.85,
|
224 |
+
edgecolor="black",
|
225 |
+
linewidth=0.7,
|
226 |
+
)
|
227 |
+
ax.set_xlabel(metric_labels.get(name, name), fontsize=11)
|
228 |
+
if idx == 0:
|
229 |
+
ax.set_ylabel("Count", fontsize=10)
|
230 |
+
# Draw limits and collect legend handles only once
|
231 |
+
if name in limits:
|
232 |
+
lims = limits[name]
|
233 |
+
if len(legend_handles) == 0:
|
234 |
+
# Only add legend handles for the first metric
|
235 |
+
min_handle = mlines.Line2D([], [], **min_style, label="min score")
|
236 |
+
max_handle = mlines.Line2D([], [], **max_style, label="max score")
|
237 |
+
legend_handles.extend([min_handle, max_handle])
|
238 |
+
if len(lims) > 0:
|
239 |
+
ax.axvline(lims[0], **min_style)
|
240 |
+
if len(lims) > 1:
|
241 |
+
ax.axvline(lims[1], **max_style)
|
242 |
+
ax.spines["top"].set_visible(False)
|
243 |
+
ax.spines["right"].set_visible(False)
|
244 |
+
ax.tick_params(axis="both", which="major", labelsize=9)
|
245 |
+
# Place legend above all subplots
|
246 |
+
fig.legend(
|
247 |
+
handles=legend_handles,
|
248 |
+
loc="upper center",
|
249 |
+
ncol=2,
|
250 |
+
fontsize=10,
|
251 |
+
frameon=False,
|
252 |
+
bbox_to_anchor=(0.5, 1.02),
|
253 |
+
)
|
254 |
+
return fig
|