Spaces:
Build error
Build error
fadindashfr
commited on
Commit
·
4c5329d
1
Parent(s):
00edf39
initial commit all file
Browse files- .gitattributes +1 -0
- Description.md +22 -0
- app.py +89 -0
- configs/evaluate.json +75 -0
- configs/inference.json +116 -0
- configs/logging.conf +21 -0
- configs/metadata.json +86 -0
- configs/multi_gpu_evaluate.json +32 -0
- configs/multi_gpu_train.json +41 -0
- configs/train.json +350 -0
- figures/architecture.png +0 -0
- models/model.pt +3 -0
- requirements.txt +2 -0
- sample_data/Images/test_11_2_0628.png +0 -0
- sample_data/Images/test_12_3_0292.png +0 -0
- sample_data/Images/test_14_3_0433.png +0 -0
- sample_data/Images/test_14_4_0544.png +0 -0
- sample_data/Images/test_9_4_0019.png +0 -0
- sample_data/Images/test_9_4_0149.png +0 -0
- sample_data/Images/train_1_1_0095.png +0 -0
- sample_data/Images/train_1_3_0020.png +0 -0
- sample_data/Labels/test_11_2_0628.png +0 -0
- sample_data/Labels/test_12_3_0292.png +0 -0
- sample_data/Labels/test_14_3_0433.png +0 -0
- sample_data/Labels/test_14_4_0544.png +0 -0
- sample_data/Labels/test_9_4_0019.png +0 -0
- sample_data/Labels/test_9_4_0149.png +0 -0
- sample_data/Labels/train_1_1_0095.png +0 -0
- sample_data/Labels/train_1_3_0020.png +0 -0
.gitattributes
CHANGED
@@ -32,3 +32,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
35 |
+
*.ts filter=lfs diff=lfs merge=lfs -text
|
Description.md
ADDED
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
## Overview
|
2 |
+
Nuclei classification within Haematoxylin & Eosi stained histology images. Classifying nuclei cells as the following types:
|
3 |
+
- Other
|
4 |
+
- Inflammatory
|
5 |
+
- Epithelial
|
6 |
+
- Spindle-Shaped
|
7 |
+
|
8 |
+
References: https://doi.org/10.1016/j.media.2019.101563
|
9 |
+
|
10 |
+
## Dataset
|
11 |
+
The model is trained with Colorectal Nuclear Segmentation and Phenotypes (CoNSeP) dataset https://warwick.ac.uk/fac/cross_fac/tia/data/hovernet. Images were extracted from 16 colorectal adenocarcinoma (CRA) WSIs.
|
12 |
+
|
13 |
+
- Target: Nuclei
|
14 |
+
- Task: Nuclei Cells Class Classification
|
15 |
+
- Modality: Image (RGB)
|
16 |
+
|
17 |
+
## Model Architecture
|
18 |
+
The model is trained using DenseNet121 over CoNSep dataset.
|
19 |
+
![alt text](file/figures/architecture.png)
|
20 |
+
|
21 |
+
## Demo
|
22 |
+
Please select or upload a nuclei histology image and label image to see Nuclei Cells Classification capabilities of this model
|
app.py
ADDED
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import torch
|
2 |
+
from monai.bundle import ConfigParser
|
3 |
+
import gradio as gr
|
4 |
+
|
5 |
+
parser = ConfigParser() # load configuration files that specify various parameters for running the MONAI workflow.
|
6 |
+
parser.read_config(f="configs/inference.json") # read the config from specified JSON file
|
7 |
+
parser.read_meta(f="configs/metadata.json") # read the metadata from specified JSON file
|
8 |
+
|
9 |
+
inference = parser.get_parsed_content("inferer")
|
10 |
+
network = parser.get_parsed_content("network_def")
|
11 |
+
preprocess = parser.get_parsed_content("preprocessing")
|
12 |
+
state_dict = torch.load("models/model.pt")
|
13 |
+
network.load_state_dict(state_dict, strict=True) # Loads a model’s parameter dictionary
|
14 |
+
class_names = {
|
15 |
+
0: "Other",
|
16 |
+
1: "Inflammatory",
|
17 |
+
2: "Epithelial",
|
18 |
+
3: "Spindle-Shaped",
|
19 |
+
}
|
20 |
+
|
21 |
+
def classify_image(image_file, label_file):
|
22 |
+
data = {"image":image_file, "label":label_file}
|
23 |
+
batch = preprocess(data)
|
24 |
+
network.eval()
|
25 |
+
with torch.no_grad():
|
26 |
+
pred = inference(batch['image'].unsqueeze(dim=0), network) # expect 4 channels input (3 RGB, 1 Label mask)
|
27 |
+
prob = pred.softmax(-1).detach().cpu().numpy()[0]
|
28 |
+
confidences = {class_names[i]: float(prob[i]) for i in range(len(class_names))}
|
29 |
+
return confidences
|
30 |
+
|
31 |
+
example_files1 = [
|
32 |
+
[r'sample_data\Images\test_11_2_0628.png',
|
33 |
+
r'sample_data\Labels\test_11_2_0628.png'],
|
34 |
+
[r'sample_data\Images\test_9_4_0149.png',
|
35 |
+
r'sample_data\Labels\test_9_4_0149.png'],
|
36 |
+
[r'sample_data\Images\test_12_3_0292.png',
|
37 |
+
r'sample_data\Labels\test_12_3_0292.png'],
|
38 |
+
[r'sample_data\Images\test_9_4_0019.png',
|
39 |
+
r'sample_data\Labels\test_9_4_0019.png']
|
40 |
+
]
|
41 |
+
|
42 |
+
example_files2 = [
|
43 |
+
[r'sample_data\Images\test_14_3_0433.png',
|
44 |
+
r'sample_data\Labels\test_14_3_0433.png'],
|
45 |
+
[r'sample_data\Images\test_14_4_0544.png',
|
46 |
+
r'sample_data\Labels\test_14_4_0544.png'],
|
47 |
+
[r'sample_data\Images\train_1_1_0095.png',
|
48 |
+
r'sample_data\Labels\train_1_1_0095.png'],
|
49 |
+
[r'sample_data\Images\train_1_3_0020.png',
|
50 |
+
r'sample_data\Labels\train_1_3_0020.png'],
|
51 |
+
]
|
52 |
+
|
53 |
+
with open('Description.md','r') as file:
|
54 |
+
markdown_content = file.read()
|
55 |
+
with gr.Blocks() as app:
|
56 |
+
gr.Markdown("# Pathology Nuclei Classification")
|
57 |
+
gr.Markdown(markdown_content)
|
58 |
+
with gr.Row():
|
59 |
+
with gr.Column():
|
60 |
+
with gr.Row():
|
61 |
+
inp_img = gr.Image(type="filepath", image_mode="RGB")
|
62 |
+
label_img = gr.Image(type="filepath", image_mode="L")
|
63 |
+
with gr.Row():
|
64 |
+
process_btn = gr.Button(value="Process")
|
65 |
+
clear_btn = gr.Button(value="Clear")
|
66 |
+
out_txt = gr.Label(label="Probabilities", num_top_classes=4)
|
67 |
+
|
68 |
+
process_btn.click(fn=classify_image, inputs=[inp_img, label_img], outputs=out_txt)
|
69 |
+
clear_btn.click(lambda:(
|
70 |
+
gr.update(value=None),
|
71 |
+
gr.update(value=None),
|
72 |
+
gr.update(value=None)
|
73 |
+
),
|
74 |
+
inputs=None,
|
75 |
+
outputs=[inp_img, label_img,out_txt]
|
76 |
+
)
|
77 |
+
|
78 |
+
gr.Markdown("## Image Examples")
|
79 |
+
with gr.Row():
|
80 |
+
for file in example_files1:
|
81 |
+
gr.Examples(
|
82 |
+
[file], inputs=[inp_img, label_img]
|
83 |
+
)
|
84 |
+
with gr.Row():
|
85 |
+
for file in example_files2:
|
86 |
+
gr.Examples(
|
87 |
+
[file], inputs=[inp_img, label_img]
|
88 |
+
)
|
89 |
+
app.launch()
|
configs/evaluate.json
ADDED
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"validate#dataset#cache_rate": 0,
|
3 |
+
"validate#postprocessing": {
|
4 |
+
"_target_": "Compose",
|
5 |
+
"transforms": [
|
6 |
+
{
|
7 |
+
"_target_": "Activationsd",
|
8 |
+
"keys": "pred",
|
9 |
+
"softmax": true
|
10 |
+
},
|
11 |
+
{
|
12 |
+
"_target_": "AsDiscreted",
|
13 |
+
"keys": [
|
14 |
+
"pred",
|
15 |
+
"label"
|
16 |
+
],
|
17 |
+
"argmax": [
|
18 |
+
true,
|
19 |
+
false
|
20 |
+
],
|
21 |
+
"to_onehot": 4
|
22 |
+
},
|
23 |
+
{
|
24 |
+
"_target_": "ToTensord",
|
25 |
+
"keys": [
|
26 |
+
"pred",
|
27 |
+
"label"
|
28 |
+
],
|
29 |
+
"device": "@device"
|
30 |
+
},
|
31 |
+
{
|
32 |
+
"_target_": "SaveImaged",
|
33 |
+
"_disabled_": true,
|
34 |
+
"keys": "pred",
|
35 |
+
"meta_keys": "pred_meta_dict",
|
36 |
+
"output_dir": "@output_dir",
|
37 |
+
"output_ext": ".json"
|
38 |
+
}
|
39 |
+
]
|
40 |
+
},
|
41 |
+
"validate#handlers": [
|
42 |
+
{
|
43 |
+
"_target_": "CheckpointLoader",
|
44 |
+
"load_path": "$@ckpt_dir + '/model.pt'",
|
45 |
+
"load_dict": {
|
46 |
+
"model": "@network"
|
47 |
+
}
|
48 |
+
},
|
49 |
+
{
|
50 |
+
"_target_": "StatsHandler",
|
51 |
+
"iteration_log": false
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"_target_": "MetricsSaver",
|
55 |
+
"save_dir": "@output_dir",
|
56 |
+
"metrics": [
|
57 |
+
"val_f1",
|
58 |
+
"val_accuracy"
|
59 |
+
],
|
60 |
+
"metric_details": [
|
61 |
+
"val_f1"
|
62 |
+
],
|
63 |
+
"batch_transform": "$monai.handlers.from_engine(['image_meta_dict'])",
|
64 |
+
"summary_ops": "*"
|
65 |
+
}
|
66 |
+
],
|
67 |
+
"evaluating": [
|
68 |
+
"$import sys",
|
69 |
+
"$sys.path.append(@bundle_root)",
|
70 |
+
"$setattr(torch.backends.cudnn, 'benchmark', True)",
|
71 |
+
"$import scripts",
|
72 |
+
"$monai.data.register_writer('json', scripts.ClassificationWriter)",
|
73 |
+
"$@validate#evaluator.run()"
|
74 |
+
]
|
75 |
+
}
|
configs/inference.json
ADDED
@@ -0,0 +1,116 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"imports": [
|
3 |
+
"$import glob",
|
4 |
+
"$import json",
|
5 |
+
"$import pathlib",
|
6 |
+
"$import os"
|
7 |
+
],
|
8 |
+
"bundle_root": "/workspace/data/pathology_nuclei_classification",
|
9 |
+
"output_dir": "$@bundle_root + '/eval'",
|
10 |
+
"dataset_dir": "/workspace/data/CoNSePNuclei",
|
11 |
+
"images": "$list(sorted(glob.glob(@dataset_dir + '/Test/Images/*.png')))[:1]",
|
12 |
+
"labels": "$list(sorted(glob.glob(@dataset_dir + '/Test/Labels/*.png')))[:1]",
|
13 |
+
"input_data": "$[{'image': i, 'label': l} for i,l in zip(@images, @labels)]",
|
14 |
+
"device": "$torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')",
|
15 |
+
"network_def": {
|
16 |
+
"_target_": "DenseNet121",
|
17 |
+
"spatial_dims": 2,
|
18 |
+
"in_channels": 4,
|
19 |
+
"out_channels": 4
|
20 |
+
},
|
21 |
+
"network": "$@network_def.to(@device)",
|
22 |
+
"preprocessing": {
|
23 |
+
"_target_": "Compose",
|
24 |
+
"transforms": [
|
25 |
+
{
|
26 |
+
"_target_": "LoadImaged",
|
27 |
+
"keys": [
|
28 |
+
"image",
|
29 |
+
"label"
|
30 |
+
],
|
31 |
+
"dtype": "uint8"
|
32 |
+
},
|
33 |
+
{
|
34 |
+
"_target_": "EnsureChannelFirstd",
|
35 |
+
"keys": [
|
36 |
+
"image",
|
37 |
+
"label"
|
38 |
+
]
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"_target_": "ScaleIntensityRanged",
|
42 |
+
"keys": "image",
|
43 |
+
"a_min": 0.0,
|
44 |
+
"a_max": 255.0,
|
45 |
+
"b_min": -1.0,
|
46 |
+
"b_max": 1.0
|
47 |
+
},
|
48 |
+
{
|
49 |
+
"_target_": "AddLabelAsGuidanced",
|
50 |
+
"keys": "image",
|
51 |
+
"source": "label"
|
52 |
+
}
|
53 |
+
]
|
54 |
+
},
|
55 |
+
"dataset": {
|
56 |
+
"_target_": "Dataset",
|
57 |
+
"data": "@input_data",
|
58 |
+
"transform": "@preprocessing"
|
59 |
+
},
|
60 |
+
"dataloader": {
|
61 |
+
"_target_": "DataLoader",
|
62 |
+
"dataset": "@dataset",
|
63 |
+
"batch_size": 1,
|
64 |
+
"shuffle": false,
|
65 |
+
"num_workers": 4
|
66 |
+
},
|
67 |
+
"inferer": {
|
68 |
+
"_target_": "SimpleInferer"
|
69 |
+
},
|
70 |
+
"postprocessing": {
|
71 |
+
"_target_": "Compose",
|
72 |
+
"transforms": [
|
73 |
+
{
|
74 |
+
"_target_": "Activationsd",
|
75 |
+
"keys": "pred",
|
76 |
+
"softmax": true
|
77 |
+
},
|
78 |
+
{
|
79 |
+
"_target_": "SaveImaged",
|
80 |
+
"keys": "pred",
|
81 |
+
"meta_keys": "pred_meta_dict",
|
82 |
+
"output_dir": "@output_dir",
|
83 |
+
"output_ext": ".json"
|
84 |
+
}
|
85 |
+
]
|
86 |
+
},
|
87 |
+
"handlers": [
|
88 |
+
{
|
89 |
+
"_target_": "CheckpointLoader",
|
90 |
+
"load_path": "$@bundle_root + '/models/model.pt'",
|
91 |
+
"load_dict": {
|
92 |
+
"model": "@network"
|
93 |
+
}
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"_target_": "StatsHandler",
|
97 |
+
"iteration_log": false
|
98 |
+
}
|
99 |
+
],
|
100 |
+
"evaluator": {
|
101 |
+
"_target_": "SupervisedEvaluator",
|
102 |
+
"device": "@device",
|
103 |
+
"val_data_loader": "@dataloader",
|
104 |
+
"network": "@network",
|
105 |
+
"inferer": "@inferer",
|
106 |
+
"postprocessing": "@postprocessing",
|
107 |
+
"val_handlers": "@handlers",
|
108 |
+
"amp": true
|
109 |
+
},
|
110 |
+
"evaluating": [
|
111 |
+
"$setattr(torch.backends.cudnn, 'benchmark', True)",
|
112 |
+
"$import scripts",
|
113 |
+
"$monai.data.register_writer('json', scripts.ClassificationWriter)",
|
114 |
+
"$@evaluator.run()"
|
115 |
+
]
|
116 |
+
}
|
configs/logging.conf
ADDED
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
[loggers]
|
2 |
+
keys=root
|
3 |
+
|
4 |
+
[handlers]
|
5 |
+
keys=consoleHandler
|
6 |
+
|
7 |
+
[formatters]
|
8 |
+
keys=fullFormatter
|
9 |
+
|
10 |
+
[logger_root]
|
11 |
+
level=INFO
|
12 |
+
handlers=consoleHandler
|
13 |
+
|
14 |
+
[handler_consoleHandler]
|
15 |
+
class=StreamHandler
|
16 |
+
level=INFO
|
17 |
+
formatter=fullFormatter
|
18 |
+
args=(sys.stdout,)
|
19 |
+
|
20 |
+
[formatter_fullFormatter]
|
21 |
+
format=%(asctime)s - %(name)s - %(levelname)s - %(message)s
|
configs/metadata.json
ADDED
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"schema": "https://github.com/Project-MONAI/MONAI-extra-test-data/releases/download/0.8.1/meta_schema_20220324.json",
|
3 |
+
"version": "0.0.5",
|
4 |
+
"changelog": {
|
5 |
+
"0.0.5": "add name tag",
|
6 |
+
"0.0.4": "Fix evaluation",
|
7 |
+
"0.0.3": "Update to use MONAI 1.1.0",
|
8 |
+
"0.0.2": "Update The Torch Vision Transform",
|
9 |
+
"0.0.1": "initialize the model package structure"
|
10 |
+
},
|
11 |
+
"monai_version": "1.1.0",
|
12 |
+
"pytorch_version": "1.13.0",
|
13 |
+
"numpy_version": "1.21.2",
|
14 |
+
"optional_packages_version": {
|
15 |
+
"nibabel": "4.0.1",
|
16 |
+
"pytorch-ignite": "0.4.9"
|
17 |
+
},
|
18 |
+
"name": "Pathology nuclei classification",
|
19 |
+
"task": "Pathology Nuclei classification",
|
20 |
+
"description": "A pre-trained model for Nuclei Classification within Haematoxylin & Eosin stained histology images",
|
21 |
+
"authors": "MONAI team",
|
22 |
+
"copyright": "Copyright (c) MONAI Consortium",
|
23 |
+
"data_source": "consep_dataset.zip from https://warwick.ac.uk/fac/cross_fac/tia/data/hovernet",
|
24 |
+
"data_type": "png",
|
25 |
+
"image_classes": "RGB channel data, intensity scaled to [0, 1]",
|
26 |
+
"label_classes": "single channel data",
|
27 |
+
"pred_classes": "4 channels OneHot data, channel 0 is Other, channel 1 is Inflammatory, channel 2 is Epithelial, channel 3 is Spindle-Shaped",
|
28 |
+
"eval_metrics": {
|
29 |
+
"f1_score": 0.85
|
30 |
+
},
|
31 |
+
"intended_use": "This is an example, not to be used for diagnostic purposes",
|
32 |
+
"references": [
|
33 |
+
"S. Graham, Q. D. Vu, S. E. A. Raza, A. Azam, Y-W. Tsang, J. T. Kwak and N. Rajpoot. \"HoVer-Net: Simultaneous Segmentation and Classification of Nuclei in Multi-Tissue Histology Images.\" Medical Image Analysis, Sept. 2019. https://doi.org/10.1016/j.media.2019.101563"
|
34 |
+
],
|
35 |
+
"network_data_format": {
|
36 |
+
"inputs": {
|
37 |
+
"image": {
|
38 |
+
"type": "magnitude",
|
39 |
+
"format": "RGB",
|
40 |
+
"modality": "regular",
|
41 |
+
"num_channels": 4,
|
42 |
+
"spatial_shape": [
|
43 |
+
128,
|
44 |
+
128
|
45 |
+
],
|
46 |
+
"dtype": "float32",
|
47 |
+
"value_range": [
|
48 |
+
0,
|
49 |
+
1
|
50 |
+
],
|
51 |
+
"is_patch_data": false,
|
52 |
+
"channel_def": {
|
53 |
+
"0": "R",
|
54 |
+
"1": "G",
|
55 |
+
"2": "B",
|
56 |
+
"3": "Mask"
|
57 |
+
}
|
58 |
+
}
|
59 |
+
},
|
60 |
+
"outputs": {
|
61 |
+
"pred": {
|
62 |
+
"type": "probabilities",
|
63 |
+
"format": "classes",
|
64 |
+
"num_channels": 4,
|
65 |
+
"spatial_shape": [
|
66 |
+
1,
|
67 |
+
4
|
68 |
+
],
|
69 |
+
"dtype": "float32",
|
70 |
+
"value_range": [
|
71 |
+
0,
|
72 |
+
1,
|
73 |
+
2,
|
74 |
+
3
|
75 |
+
],
|
76 |
+
"is_patch_data": false,
|
77 |
+
"channel_def": {
|
78 |
+
"0": "Other",
|
79 |
+
"1": "Inflammatory",
|
80 |
+
"2": "Epithelial",
|
81 |
+
"3": "Spindle-Shaped"
|
82 |
+
}
|
83 |
+
}
|
84 |
+
}
|
85 |
+
}
|
86 |
+
}
|
configs/multi_gpu_evaluate.json
ADDED
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"device": "$torch.device(f'cuda:{dist.get_rank()}')",
|
3 |
+
"network": {
|
4 |
+
"_target_": "torch.nn.parallel.DistributedDataParallel",
|
5 |
+
"module": "$@network_def.to(@device)",
|
6 |
+
"device_ids": [
|
7 |
+
"@device"
|
8 |
+
]
|
9 |
+
},
|
10 |
+
"validate#sampler": {
|
11 |
+
"_target_": "DistributedSampler",
|
12 |
+
"dataset": "@validate#dataset",
|
13 |
+
"even_divisible": false,
|
14 |
+
"shuffle": false
|
15 |
+
},
|
16 |
+
"validate#dataloader#sampler": "@validate#sampler",
|
17 |
+
"validate#handlers#1#_disabled_": "$dist.get_rank() > 0",
|
18 |
+
"evaluating": [
|
19 |
+
"$import sys",
|
20 |
+
"$sys.path.append(@bundle_root)",
|
21 |
+
"$import torch.distributed as dist",
|
22 |
+
"$dist.init_process_group(backend='nccl')",
|
23 |
+
"$torch.cuda.set_device(@device)",
|
24 |
+
"$setattr(torch.backends.cudnn, 'benchmark', True)",
|
25 |
+
"$import logging",
|
26 |
+
"$@validate#evaluator.logger.setLevel(logging.WARNING if dist.get_rank() > 0 else logging.INFO)",
|
27 |
+
"$import scripts",
|
28 |
+
"$monai.data.register_writer('json', scripts.ClassificationWriter)",
|
29 |
+
"$@validate#evaluator.run()",
|
30 |
+
"$dist.destroy_process_group()"
|
31 |
+
]
|
32 |
+
}
|
configs/multi_gpu_train.json
ADDED
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"device": "$torch.device(f'cuda:{dist.get_rank()}')",
|
3 |
+
"network": {
|
4 |
+
"_target_": "torch.nn.parallel.DistributedDataParallel",
|
5 |
+
"module": "$@network_def.to(@device)",
|
6 |
+
"device_ids": [
|
7 |
+
"@device"
|
8 |
+
]
|
9 |
+
},
|
10 |
+
"train#sampler": {
|
11 |
+
"_target_": "DistributedSampler",
|
12 |
+
"dataset": "@train#dataset",
|
13 |
+
"even_divisible": true,
|
14 |
+
"shuffle": true
|
15 |
+
},
|
16 |
+
"train#dataloader#sampler": "@train#sampler",
|
17 |
+
"train#dataloader#shuffle": false,
|
18 |
+
"train#trainer#train_handlers": "$@train#handlers[: -2 if dist.get_rank() > 0 else None]",
|
19 |
+
"validate#sampler": {
|
20 |
+
"_target_": "DistributedSampler",
|
21 |
+
"dataset": "@validate#dataset",
|
22 |
+
"even_divisible": false,
|
23 |
+
"shuffle": false
|
24 |
+
},
|
25 |
+
"validate#dataloader#sampler": "@validate#sampler",
|
26 |
+
"validate#evaluator#val_handlers": "$None if dist.get_rank() > 0 else @validate#handlers",
|
27 |
+
"training": [
|
28 |
+
"$import sys",
|
29 |
+
"$sys.path.append(@bundle_root)",
|
30 |
+
"$import torch.distributed as dist",
|
31 |
+
"$dist.init_process_group(backend='nccl')",
|
32 |
+
"$torch.cuda.set_device(@device)",
|
33 |
+
"$monai.utils.set_determinism(seed=123)",
|
34 |
+
"$setattr(torch.backends.cudnn, 'benchmark', True)",
|
35 |
+
"$import logging",
|
36 |
+
"$@train#trainer.logger.setLevel(logging.WARNING if dist.get_rank() > 0 else logging.INFO)",
|
37 |
+
"$@validate#evaluator.logger.setLevel(logging.WARNING if dist.get_rank() > 0 else logging.INFO)",
|
38 |
+
"$@train#trainer.run()",
|
39 |
+
"$dist.destroy_process_group()"
|
40 |
+
]
|
41 |
+
}
|
configs/train.json
ADDED
@@ -0,0 +1,350 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"imports": [
|
3 |
+
"$import glob",
|
4 |
+
"$import ignite",
|
5 |
+
"$import json",
|
6 |
+
"$import pathlib",
|
7 |
+
"$import os"
|
8 |
+
],
|
9 |
+
"bundle_root": "/workspace/data/pathology_nuclei_classification",
|
10 |
+
"ckpt_dir": "$@bundle_root + '/models'",
|
11 |
+
"output_dir": "$@bundle_root + '/eval'",
|
12 |
+
"dataset_dir": "/workspace/data/CoNSePNuclei",
|
13 |
+
"dataset_json": "$@dataset_dir + '/dataset.json'",
|
14 |
+
"train_datalist": "$json.loads(pathlib.Path(@dataset_json).read_text())['training']",
|
15 |
+
"val_datalist": "$json.loads(pathlib.Path(@dataset_json).read_text())['validation']",
|
16 |
+
"val_interval": 1,
|
17 |
+
"device": "$torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')",
|
18 |
+
"network_def": {
|
19 |
+
"_target_": "DenseNet121",
|
20 |
+
"spatial_dims": 2,
|
21 |
+
"in_channels": 4,
|
22 |
+
"out_channels": 4
|
23 |
+
},
|
24 |
+
"network": "$@network_def.to(@device)",
|
25 |
+
"loss": {
|
26 |
+
"_target_": "torch.nn.CrossEntropyLoss"
|
27 |
+
},
|
28 |
+
"optimizer": {
|
29 |
+
"_target_": "torch.optim.Adam",
|
30 |
+
"params": "$@network.parameters()",
|
31 |
+
"lr": 0.0001
|
32 |
+
},
|
33 |
+
"max_epochs": 50,
|
34 |
+
"train": {
|
35 |
+
"preprocessing": {
|
36 |
+
"_target_": "Compose",
|
37 |
+
"transforms": [
|
38 |
+
{
|
39 |
+
"_target_": "LoadImaged",
|
40 |
+
"keys": [
|
41 |
+
"image",
|
42 |
+
"label"
|
43 |
+
],
|
44 |
+
"dtype": "uint8"
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"_target_": "EnsureChannelFirstd",
|
48 |
+
"keys": [
|
49 |
+
"image",
|
50 |
+
"label"
|
51 |
+
]
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"_target_": "SplitLabeld",
|
55 |
+
"keys": "label",
|
56 |
+
"mask_value": "",
|
57 |
+
"others_value": 255,
|
58 |
+
"to_binary_mask": false
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"_target_": "RandTorchVisiond",
|
62 |
+
"keys": "image",
|
63 |
+
"name": "ColorJitter",
|
64 |
+
"brightness": 0.25,
|
65 |
+
"contrast": 0.75,
|
66 |
+
"saturation": 0.25,
|
67 |
+
"hue": 0.04
|
68 |
+
},
|
69 |
+
{
|
70 |
+
"_target_": "RandFlipd",
|
71 |
+
"keys": [
|
72 |
+
"image",
|
73 |
+
"label",
|
74 |
+
"others"
|
75 |
+
],
|
76 |
+
"prob": 0.5
|
77 |
+
},
|
78 |
+
{
|
79 |
+
"_target_": "RandRotate90d",
|
80 |
+
"keys": [
|
81 |
+
"image",
|
82 |
+
"label",
|
83 |
+
"others"
|
84 |
+
],
|
85 |
+
"prob": 0.5
|
86 |
+
},
|
87 |
+
{
|
88 |
+
"_target_": "ScaleIntensityRanged",
|
89 |
+
"keys": "image",
|
90 |
+
"a_min": 0.0,
|
91 |
+
"a_max": 255.0,
|
92 |
+
"b_min": -1.0,
|
93 |
+
"b_max": 1.0
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"_target_": "AddLabelAsGuidanced",
|
97 |
+
"keys": "image",
|
98 |
+
"source": "label"
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"_target_": "SetLabelClassd",
|
102 |
+
"keys": "label",
|
103 |
+
"offset": -1
|
104 |
+
},
|
105 |
+
{
|
106 |
+
"_target_": "SelectItemsd",
|
107 |
+
"keys": [
|
108 |
+
"image",
|
109 |
+
"label"
|
110 |
+
]
|
111 |
+
}
|
112 |
+
]
|
113 |
+
},
|
114 |
+
"dataset": {
|
115 |
+
"_target_": "CacheDataset",
|
116 |
+
"data": "@train_datalist",
|
117 |
+
"transform": "@train#preprocessing",
|
118 |
+
"cache_rate": 1.0,
|
119 |
+
"num_workers": 4
|
120 |
+
},
|
121 |
+
"dataloader": {
|
122 |
+
"_target_": "DataLoader",
|
123 |
+
"dataset": "@train#dataset",
|
124 |
+
"batch_size": 64,
|
125 |
+
"shuffle": true,
|
126 |
+
"num_workers": 4
|
127 |
+
},
|
128 |
+
"inferer": {
|
129 |
+
"_target_": "SimpleInferer"
|
130 |
+
},
|
131 |
+
"postprocessing": {
|
132 |
+
"_target_": "Compose",
|
133 |
+
"transforms": [
|
134 |
+
{
|
135 |
+
"_target_": "Activationsd",
|
136 |
+
"keys": "pred",
|
137 |
+
"softmax": true
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"_target_": "AsDiscreted",
|
141 |
+
"keys": [
|
142 |
+
"pred",
|
143 |
+
"label"
|
144 |
+
],
|
145 |
+
"argmax": [
|
146 |
+
true,
|
147 |
+
false
|
148 |
+
],
|
149 |
+
"to_onehot": 4
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"_target_": "ToTensord",
|
153 |
+
"keys": [
|
154 |
+
"pred",
|
155 |
+
"label"
|
156 |
+
],
|
157 |
+
"device": "@device"
|
158 |
+
}
|
159 |
+
]
|
160 |
+
},
|
161 |
+
"handlers": [
|
162 |
+
{
|
163 |
+
"_target_": "ValidationHandler",
|
164 |
+
"validator": "@validate#evaluator",
|
165 |
+
"epoch_level": true,
|
166 |
+
"interval": "@val_interval"
|
167 |
+
},
|
168 |
+
{
|
169 |
+
"_target_": "StatsHandler",
|
170 |
+
"tag_name": "train_loss",
|
171 |
+
"output_transform": "$monai.handlers.from_engine(['loss'], first=True)"
|
172 |
+
},
|
173 |
+
{
|
174 |
+
"_target_": "TensorBoardStatsHandler",
|
175 |
+
"log_dir": "@output_dir",
|
176 |
+
"tag_name": "train_loss",
|
177 |
+
"output_transform": "$monai.handlers.from_engine(['loss'], first=True)"
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"_target_": "scripts.TensorBoardImageHandler",
|
181 |
+
"class_names": {
|
182 |
+
"0": "Other",
|
183 |
+
"1": "Inflammatory",
|
184 |
+
"2": "Epithelial",
|
185 |
+
"3": "Spindle-Shaped"
|
186 |
+
},
|
187 |
+
"log_dir": "@output_dir",
|
188 |
+
"batch_limit": 4,
|
189 |
+
"tag_name": "train"
|
190 |
+
}
|
191 |
+
],
|
192 |
+
"key_metric": {
|
193 |
+
"train_f1": {
|
194 |
+
"_target_": "ConfusionMatrix",
|
195 |
+
"metric_name": "f1 score",
|
196 |
+
"output_transform": "$monai.handlers.from_engine(['pred', 'label'])"
|
197 |
+
}
|
198 |
+
},
|
199 |
+
"trainer": {
|
200 |
+
"_target_": "SupervisedTrainer",
|
201 |
+
"max_epochs": "@max_epochs",
|
202 |
+
"device": "@device",
|
203 |
+
"train_data_loader": "@train#dataloader",
|
204 |
+
"network": "@network",
|
205 |
+
"loss_function": "@loss",
|
206 |
+
"optimizer": "@optimizer",
|
207 |
+
"inferer": "@train#inferer",
|
208 |
+
"postprocessing": "@train#postprocessing",
|
209 |
+
"key_train_metric": "@train#key_metric",
|
210 |
+
"train_handlers": "@train#handlers",
|
211 |
+
"amp": true
|
212 |
+
}
|
213 |
+
},
|
214 |
+
"validate": {
|
215 |
+
"preprocessing": {
|
216 |
+
"_target_": "Compose",
|
217 |
+
"transforms": [
|
218 |
+
{
|
219 |
+
"_target_": "LoadImaged",
|
220 |
+
"keys": [
|
221 |
+
"image",
|
222 |
+
"label"
|
223 |
+
],
|
224 |
+
"dtype": "uint8"
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"_target_": "EnsureChannelFirstd",
|
228 |
+
"keys": [
|
229 |
+
"image",
|
230 |
+
"label"
|
231 |
+
]
|
232 |
+
},
|
233 |
+
{
|
234 |
+
"_target_": "SplitLabeld",
|
235 |
+
"keys": "label",
|
236 |
+
"mask_value": "",
|
237 |
+
"others_value": 255,
|
238 |
+
"to_binary_mask": false
|
239 |
+
},
|
240 |
+
{
|
241 |
+
"_target_": "ScaleIntensityRanged",
|
242 |
+
"keys": "image",
|
243 |
+
"a_min": 0.0,
|
244 |
+
"a_max": 255.0,
|
245 |
+
"b_min": -1.0,
|
246 |
+
"b_max": 1.0
|
247 |
+
},
|
248 |
+
{
|
249 |
+
"_target_": "AddLabelAsGuidanced",
|
250 |
+
"keys": "image",
|
251 |
+
"source": "label"
|
252 |
+
},
|
253 |
+
{
|
254 |
+
"_target_": "SetLabelClassd",
|
255 |
+
"keys": "label",
|
256 |
+
"offset": -1
|
257 |
+
},
|
258 |
+
{
|
259 |
+
"_target_": "SelectItemsd",
|
260 |
+
"keys": [
|
261 |
+
"image",
|
262 |
+
"label",
|
263 |
+
"image_meta_dict"
|
264 |
+
]
|
265 |
+
}
|
266 |
+
]
|
267 |
+
},
|
268 |
+
"dataset": {
|
269 |
+
"_target_": "CacheDataset",
|
270 |
+
"data": "@val_datalist",
|
271 |
+
"transform": "@validate#preprocessing",
|
272 |
+
"cache_rate": 1.0
|
273 |
+
},
|
274 |
+
"dataloader": {
|
275 |
+
"_target_": "DataLoader",
|
276 |
+
"dataset": "@validate#dataset",
|
277 |
+
"batch_size": 64,
|
278 |
+
"shuffle": false,
|
279 |
+
"num_workers": 4
|
280 |
+
},
|
281 |
+
"inferer": {
|
282 |
+
"_target_": "SimpleInferer"
|
283 |
+
},
|
284 |
+
"postprocessing": "%train#postprocessing",
|
285 |
+
"handlers": [
|
286 |
+
{
|
287 |
+
"_target_": "StatsHandler",
|
288 |
+
"iteration_log": false
|
289 |
+
},
|
290 |
+
{
|
291 |
+
"_target_": "TensorBoardStatsHandler",
|
292 |
+
"log_dir": "@output_dir",
|
293 |
+
"iteration_log": false
|
294 |
+
},
|
295 |
+
{
|
296 |
+
"_target_": "CheckpointSaver",
|
297 |
+
"save_dir": "@ckpt_dir",
|
298 |
+
"save_dict": {
|
299 |
+
"model": "@network"
|
300 |
+
},
|
301 |
+
"save_key_metric": true,
|
302 |
+
"key_metric_filename": "model.pt"
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"_target_": "scripts.TensorBoardImageHandler",
|
306 |
+
"class_names": {
|
307 |
+
"0": "Other",
|
308 |
+
"1": "Inflammatory",
|
309 |
+
"2": "Epithelial",
|
310 |
+
"3": "Spindle-Shaped"
|
311 |
+
},
|
312 |
+
"log_dir": "@output_dir",
|
313 |
+
"batch_limit": 8,
|
314 |
+
"tag_name": "val"
|
315 |
+
}
|
316 |
+
],
|
317 |
+
"key_metric": {
|
318 |
+
"val_f1": {
|
319 |
+
"_target_": "ConfusionMatrix",
|
320 |
+
"metric_name": "f1 score",
|
321 |
+
"output_transform": "$monai.handlers.from_engine(['pred', 'label'])"
|
322 |
+
}
|
323 |
+
},
|
324 |
+
"additional_metrics": {
|
325 |
+
"val_accuracy": {
|
326 |
+
"_target_": "ignite.metrics.Accuracy",
|
327 |
+
"output_transform": "$monai.handlers.from_engine(['pred', 'label'])"
|
328 |
+
}
|
329 |
+
},
|
330 |
+
"evaluator": {
|
331 |
+
"_target_": "SupervisedEvaluator",
|
332 |
+
"device": "@device",
|
333 |
+
"val_data_loader": "@validate#dataloader",
|
334 |
+
"network": "@network",
|
335 |
+
"inferer": "@validate#inferer",
|
336 |
+
"postprocessing": "@validate#postprocessing",
|
337 |
+
"key_val_metric": "@validate#key_metric",
|
338 |
+
"additional_metrics": "@validate#additional_metrics",
|
339 |
+
"val_handlers": "@validate#handlers",
|
340 |
+
"amp": true
|
341 |
+
}
|
342 |
+
},
|
343 |
+
"training": [
|
344 |
+
"$import sys",
|
345 |
+
"$sys.path.append(@bundle_root)",
|
346 |
+
"$monai.utils.set_determinism(seed=123)",
|
347 |
+
"$setattr(torch.backends.cudnn, 'benchmark', True)",
|
348 |
+
"$@train#trainer.run()"
|
349 |
+
]
|
350 |
+
}
|
figures/architecture.png
ADDED
models/model.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f254ae6b0318e1375d48c1e9d6056d236d4a1a32957afc4aeafba0e047c46b2b
|
3 |
+
size 28419489
|
requirements.txt
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
monai==1.1.0
|
2 |
+
gradio
|
sample_data/Images/test_11_2_0628.png
ADDED
sample_data/Images/test_12_3_0292.png
ADDED
sample_data/Images/test_14_3_0433.png
ADDED
sample_data/Images/test_14_4_0544.png
ADDED
sample_data/Images/test_9_4_0019.png
ADDED
sample_data/Images/test_9_4_0149.png
ADDED
sample_data/Images/train_1_1_0095.png
ADDED
sample_data/Images/train_1_3_0020.png
ADDED
sample_data/Labels/test_11_2_0628.png
ADDED
sample_data/Labels/test_12_3_0292.png
ADDED
sample_data/Labels/test_14_3_0433.png
ADDED
sample_data/Labels/test_14_4_0544.png
ADDED
sample_data/Labels/test_9_4_0019.png
ADDED
sample_data/Labels/test_9_4_0149.png
ADDED
sample_data/Labels/train_1_1_0095.png
ADDED
sample_data/Labels/train_1_3_0020.png
ADDED