A pre-trained model for segmenting nuclei cells with user clicks/interactions.
This model is trained using BasicUNet over ConSeP dataset.
The training dataset is from https://warwick.ac.uk/fac/cross_fac/tia/data/hovernet
wget https://warwick.ac.uk/fac/cross_fac/tia/data/hovernet/consep_dataset.zip
unzip -q consep_dataset.zip
After downloading this dataset,
python script data_process.py
from scripts
folder can be used to preprocess and generate the final dataset for training.
python scripts/data_process.py --input /path/to/data/CoNSeP --output /path/to/data/CoNSePNuclei
After generating the output files, please modify the dataset_dir
parameter specified in configs/train.json
and configs/inference.json
to reflect the output folder which contains new dataset.json.
Class values in dataset are
As part of pre-processing, the following steps are executed.
Example dataset.json
{
"training": [
{
"image": "/workspace/data/CoNSePNuclei/Train/Images/train_1_3_0001.png",
"label": "/workspace/data/CoNSePNuclei/Train/Labels/train_1_3_0001.png",
"nuclei_id": 1,
"mask_value": 3,
"centroid": [
64,
64
]
}
],
"validation": [
{
"image": "/workspace/data/CoNSePNuclei/Test/Images/test_1_3_0001.png",
"label": "/workspace/data/CoNSePNuclei/Test/Labels/test_1_3_0001.png",
"nuclei_id": 1,
"mask_value": 3,
"centroid": [
64,
64
]
}
]
}
The training was performed with the following:
If you face memory issues with CacheDataset, you can either switch to a regular Dataset class or lower the caching rate cache_rate
in the configurations within range [0, 1] to minimize the System RAM requirements.
5 channels
2 channels
This model achieves the following Dice score on the validation data provided as part of the dataset:
A graph showing the training Loss and Dice over 50 epochs.
A graph showing the validation mean Dice over 50 epochs.
In addition to the Pythonic APIs, a few command line interfaces (CLI) are provided to interact with the bundle. The CLI supports flexible use cases, such as overriding configs at runtime and predefining arguments in a file.
For more details usage instructions, visit the MONAI Bundle Configuration Page.
python -m monai.bundle run --config_file configs/train.json
Please note that if the default dataset path is not modified with the actual path in the bundle config files, you can also override it by using --dataset_dir
:
python -m monai.bundle run --config_file configs/train.json --dataset_dir <actual dataset path>
train
config to execute multi-GPU training:torchrun --standalone --nnodes=1 --nproc_per_node=2 -m monai.bundle run --config_file "['configs/train.json','configs/multi_gpu_train.json']"
Please note that the distributed training-related options depend on the actual running environment; thus, users may need to remove --standalone
, modify --nnodes
, or do some other necessary changes according to the machine used. For more details, please refer to pytorch's official tutorial.
train
config to execute evaluation with the trained model:python -m monai.bundle run --config_file "['configs/train.json','configs/evaluate.json']"
train
config and evaluate
config to execute multi-GPU evaluation:torchrun --standalone --nnodes=1 --nproc_per_node=2 -m monai.bundle run --config_file "['configs/train.json','configs/evaluate.json','configs/multi_gpu_evaluate.json']"
python -m monai.bundle run --config_file configs/inference.json
[1] Koohbanani, Navid Alemi, et al. "NuClick: a deep learning framework for interactive segmentation of microscopic images." Medical Image Analysis 65 (2020): 101771. https://arxiv.org/abs/2005.14511.
[2] 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. [doi]
[3] NuClick PyTorch Implementation
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.
By pulling and using the MONAI Toolkit container and MONAI models, you accept the terms and conditions of the NVIDIA AI Product Agreement. Register for a free evaluation license to try NVIDIA AI Enterprise on your compatible, on-premises system or on the cloud!