dungnb1333 / siim-covid19-detection Goto Github PK

View Code? Open in Web Editor NEW

172.0 8.0 58.0 9.39 MB

1st place solution for SIIM-FISABIO-RSNA COVID-19 Detection Challenge

License: MIT License

Python 54.62% Shell 1.14% Jupyter Notebook 44.18% Dockerfile 0.06%

siim-covid19-detection's Introduction

SIIM-COVID19-Detection

Source code of the 1st place solution for SIIM-FISABIO-RSNA COVID-19 Detection Challenge.

1.INSTALLATION

HARDWARE: 4 x NVIDIA TESLA V100 32G GPU, 64 vCPUs, 256GB RAM
Ubuntu 18.04.5 LTS
CUDA 10.2
Python 3.7.9
python packages are detailed separately in requirements

$ conda create -n envs python=3.7.9
$ conda activate envs
$ conda install -c conda-forge gdcm
$ pip install -r requirements.txt
$ pip install git+https://github.com/ildoonet/pytorch-gradual-warmup-lr.git
$ pip install git+https://github.com/bes-dev/mean_average_precision.git@930df3618c924b694292cc125114bad7c7f3097e

2.DATASET

2.1 SIIM COVID 19 DATASET

download competition dataset at link then extract to ./dataset/siim-covid19-detection

$ cd src/prepare
$ python dicom2image_siim.py
$ python kfold_split.py
$ python prepare_siim_annotation.py                        # effdet and yolo format
$ cp -r ../../dataset/siim-covid19-detection/images ../../dataset/lung_crop/.
$ python prepare_siim_lung_crop_annotation.py

2.2 EXTERNAL DATASET

download pneumothorax dataset at link then extract to ./dataset/external_dataset/pneumothorax/dicoms
download pneumonia dataset at link then extract to ./dataset/external_dataset/rsna-pneumonia-detection-challenge/dicoms
download vinbigdata dataset at link then extract to ./dataset/external_dataset/vinbigdata/dicoms
download chest14 dataset at link then extract to ./dataset/external_dataset/chest14/images
download chexpert high-resolution dataset at link then extract to ./dataset/external_dataset/chexpert/train
download padchest dataset at link then extract to ./dataset/external_dataset/padchest/images
_{Note: most of the images in bimcv and ricord duplicate with siim covid trainset and testset. To avoid data-leak when training, I didn't use them. You can use script src/prepare/check_bimcv_ricord_dup.py}

$ cd src/prepare
$ python dicom2image_pneumothorax.py
$ python dicom2image_pneumonia.py
$ python prepare_pneumonia_annotation.py      # effdet and yolo format
$ python dicom2image_vinbigdata.py
$ python prepare_vinbigdata.py
$ python refine_data.py                       # remove unused file in chexpert + chest14 + padchest dataset
$ python resize_padchest_pneumothorax.py

dataset structure should be ./dataset/dataset_structure.txt

3.SOLUTION SUMMARY

4.TRAIN MODEL

4.1 Classification

4.1.1 Multi head classification + segmentation

Stage1

$ cd src/classification_aux
$ bash train_chexpert_chest14.sh              #Pretrain backbone on chexpert + chest14
$ bash train_rsnapneu.sh                      #Pretrain rsna_pneumonia
$ bash train_siim.sh                          #Train siim covid19

Stage2: Generate soft-label for classification head and mask for segmentation head.
Output: soft-label in ./pseudo_csv/[source].csv and public test masks in ./prediction_mask/public_test/masks

$ bash generate_pseudo_label.sh [checkpoints_dir]

Stage3: Train model on trainset + public testset, load checkpoint from previous round

$ bash train_pseudo.sh [previous_checkpoints_dir] [new_checkpoints_dir]

Rounds of pseudo labeling (stage2) and retraining (stage3) were repeated until the score on public LB didn't improve.

For final checkpoints

$ bash generate_pseudo_label.sh checkpoints_v3
$ bash train_pseudo.sh checkpoints_v3 checkpoints_v4

For evaluation

$ CUDA_VISIBLE_DEVICES=0 python evaluate.py --cfg configs/xxx.yaml --num_tta xxx

[email protected] 4 classes: negative, typical, indeterminate, atypical

	SeR152-Unet	EB5-Deeplab	EB6-Linknet	EB7-Unet++	Ensemble
w/o TTA/8TTA	0.575/0.584	0.583/0.592	0.580/0.587	0.589/0.595	0.595/0.598

_{8TTA: (orig, center-crop 80%)x(None, hflip, vflip, hflip & vflip). In final submission, I use 4.1.2 lung detector instead of center-crop 80%}

4.1.2 Lung Detector-YoloV5

I annotated the train data(6334 images) using LabelImg and built a lung localizer. I noticed that increasing input image size improves the modeling performance and lung detector helps the model to reduce background noise.

$ cd src/detection_lung_yolov5
$ cd weights && bash download_coco_weights.sh && cd ..
$ bash train.sh

	Fold0	Fold1	Fold2	Fold3	Fold4	Average
[email protected]:0.95	0.921	0.931	0.926	0.923	0.922	0.9246
[email protected]	0.997	0.998	0.997	0.996	0.998	0.9972

4.2 Opacity Detection

Rounds of pseudo labeling (stage2) and retraining (stage3) were repeated until the score on public LB didn't improve.

4.2.1 YoloV5x6 768

Stage1:

$ cd src/detection_yolov5
$ cd weights && bash download_coco_weights.sh && cd ..
$ bash train_rsnapneu.sh          #pretrain with rsna_pneumonia
$ bash train_siim.sh              #train with siim covid19 dataset, load rsna_pneumonia checkpoint

Stage2: Generate pseudo label (boxes)

$ bash generate_pseudo_label.sh

Jump to step 4.2.4 Ensembling + Pseudo labeling

Stage3:

$ bash warmup_ext_dataset.sh      #train with pseudo labeling (public-test, padchest, pneumothorax, vin) + rsna_pneumonia
$ bash train_final.sh             #train siim covid19 boxes, load warmup checkpoint

4.2.2 EfficientDet D7 768

Stage1:

$ cd src/detection_efffdet
$ bash train_rsnapneu.sh          #pretrain with rsna_pneumonia
$ bash train_siim.sh              #train with siim covid19 dataset, load rsna_pneumonia checkpoint

Stage2: Generate pseudo label (boxes)

$ bash generate_pseudo_label.sh

Jump to step 4.2.4 Ensembling + Pseudo labeling

Stage3:

$ bash warmup_ext_dataset.sh      #train with pseudo labeling (public-test, padchest, pneumothorax, vin) + rsna_pneumonia
$ bash train_final.sh             #train siim covid19, load warmup checkpoint

4.2.3 FasterRCNN FPN 768 & 1024

Stage1: train backbone of model with chexpert + chest14 -> train model with rsna pneummonia -> train model with siim, load rsna pneumonia checkpoint

$ cd src/detection_fasterrcnn
$ CUDA_VISIBLE_DEVICES=0,1,2,3 python train_chexpert_chest14.py --steps 0 1 --cfg configs/resnet200d.yaml
$ CUDA_VISIBLE_DEVICES=0,1,2,3 python train_chexpert_chest14.py --steps 0 1 --cfg configs/resnet101d.yaml
$ CUDA_VISIBLE_DEVICES=0 python train_rsnapneu.py --cfg configs/resnet200d.yaml
$ CUDA_VISIBLE_DEVICES=0 python train_rsnapneu.py --cfg configs/resnet101d.yaml
$ CUDA_VISIBLE_DEVICES=0 python train_siim.py --cfg configs/resnet200d.yaml --folds 0 1 2 3 4 --SEED 123
$ CUDA_VISIBLE_DEVICES=0 python train_siim.py --cfg configs/resnet101d.yaml --folds 0 1 2 3 4 --SEED 123

_{Note: Change SEED if training script runs into issue related to augmentation (boundingbox area=0) and comment/uncomment the following code if training script runs into issue related to resource limit}

import resource
rlimit = resource.getrlimit(resource.RLIMIT_NOFILE)
resource.setrlimit(resource.RLIMIT_NOFILE, (8192, rlimit[1]))

Stage2: Generate pseudo label (boxes)

$ bash generate_pseudo_label.sh

Jump to step 4.2.4 Ensembling + Pseudo labeling

Stage3:

$ CUDA_VISIBLE_DEVICES=0 python warmup_ext_dataset.py --cfg configs/resnet200d.yaml
$ CUDA_VISIBLE_DEVICES=0 python warmup_ext_dataset.py --cfg configs/resnet101d.yaml
$ CUDA_VISIBLE_DEVICES=0 python train_final.py --cfg configs/resnet200d.yaml
$ CUDA_VISIBLE_DEVICES=0 python train_final.py --cfg configs/resnet101d.yaml

4.2.4 Ensembling + Pseudo labeling

Keep images that meet the conditions: negative prediction < 0.3 and maximum of (typical, indeterminate, atypical) predicion > 0.7. Then choose 2 boxes with the highest confidence as pseudo labels for each image.

_{Note: This step requires at least 128 GB of RAM}

$ cd ./src/detection_make_pseudo
$ python make_pseudo.py
$ python make_annotation.py

4.2.5 Detection Performance

	YoloV5x6 768	EffdetD7 768	F-RCNN R200 768	F-RCNN R101 1024
[email protected] TTA	0.580	0.594	0.592	0.596

5.FINAL SUBMISSION

siim-covid19-2021 Public LB: 0.658 / Private LB: 0.635
demo notebook to visualize output of models

6.AWESOME RESOURCES

Pytorch✨
PyTorch Image Models✨
Segmentation models✨
EfficientDet✨
YoloV5✨
FasterRCNN FPN✨
Albumentations✨
Weighted boxes fusion✨

siim-covid19-detection's People

Contributors

Stargazers

Watchers

siim-covid19-detection's Issues

Why is there only one category

Thank you very much for sharing your project，But I have one thing I don't understand：why is there only one category， and there are four categories of official data sets. I don't understand it，could you please answer this question，thanks.......

Not reproducable effdet results because native amp

Hello, thank you for code. But I found strange line:
https://github.com/dungnb1333/SIIM-COVID19-Detection/blob/main/src/detection_efffdet/predict_test.py#L161
If I use inference with torch native amp autocast(I didn't check with apex), result will be different every time.

Suggest to loosen the dependency on albumentations

Dear developers,

Your project SIIM-COVID19-Detection requires "albumentations==1.0.0" in its dependency. After analyzing the source code, we found that the following versions of albumentations can also be suitable without affecting your project, i.e., albumentations 1.0.1, 1.0.2, 1.0.3. Therefore, we suggest to loosen the dependency on albumentations from "albumentations==1.0.0" to "albumentations>=1.0.0,<=1.0.3" to avoid any possible conflict for importing more packages or for downstream projects that may use ddos_script.

May I pull a request to further loosen the dependency on albumentations?

By the way, could you please tell us whether such dependency analysis may be potentially helpful for maintaining dependencies easier during your development?

Details:

Your project (commit id: 982a6b1) directly uses 24 APIs from package albumentations.

albumentations.augmentations.transforms.MotionBlur.__init__, albumentations.augmentations.geometric.resize.Resize.__init__, albumentations.pytorch.transforms.ToTensorV2.__init__, albumentations.augmentations.transforms.CLAHE.__init__, albumentations.imgaug.transforms.IAASharpen.__init__, albumentations.augmentations.transforms.RandomContrast.__init__, albumentations.augmentations.geometric.rotate.Rotate.__init__, albumentations.core.composition.OneOf.__init__, albumentations.augmentations.transforms.Normalize.__init__, albumentations.augmentations.transforms.Cutout.__init__, albumentations.augmentations.transforms.RandomBrightnessContrast.__init__, albumentations.core.composition.Compose.__init__, albumentations.augmentations.crops.transforms.RandomResizedCrop.__init__, albumentations.augmentations.transforms.RandomBrightness.__init__, albumentations.augmentations.transforms.MedianBlur.__init__, albumentations.imgaug.transforms.IAAAdditiveGaussianNoise.__init__, albumentations.augmentations.geometric.resize.LongestMaxSize.__init__, albumentations.augmentations.transforms.RandomGamma.__init__, albumentations.augmentations.geometric.transforms.ShiftScaleRotate.__init__, albumentations.augmentations.transforms.Blur.__init__, albumentations.augmentations.transforms.VerticalFlip.__init__, albumentations.imgaug.transforms.IAAEmboss.__init__, albumentations.core.composition.BboxParams.__init__, albumentations.augmentations.transforms.HorizontalFlip.__init__

Beginning fromwhich, 13 functions are then indirectly called, including 12 albumentations's internal APIs and 1 outsider APIs as follows:

[/dungnb1333/SIIM-COVID19-Detection]
+--albumentations.augmentations.transforms.MotionBlur.__init__
|      +--albumentations.augmentations.transforms.Blur.__init__
|      |      +--albumentations.core.transforms_interface.BasicTransform.__init__
|      |      +--albumentations.core.transforms_interface.to_tuple
+--albumentations.augmentations.geometric.resize.Resize.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
+--albumentations.pytorch.transforms.ToTensorV2.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
+--albumentations.augmentations.transforms.CLAHE.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
|      +--albumentations.core.transforms_interface.to_tuple
+--albumentations.imgaug.transforms.IAASharpen.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
|      +--albumentations.core.transforms_interface.to_tuple
|      +--warnings.warn
+--albumentations.augmentations.transforms.RandomContrast.__init__
|      +--albumentations.augmentations.transforms.RandomBrightnessContrast.__init__
|      |      +--albumentations.core.transforms_interface.BasicTransform.__init__
|      |      +--albumentations.core.transforms_interface.to_tuple
|      +--warnings.warn
+--albumentations.augmentations.geometric.rotate.Rotate.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
|      +--albumentations.core.transforms_interface.to_tuple
+--albumentations.core.composition.OneOf.__init__
|      +--albumentations.core.composition.BaseCompose.__init__
|      |      +--albumentations.core.composition.Transforms.__init__
|      |      |      +--albumentations.core.composition.Transforms._find_dual_start_end
|      |      |      |      +--albumentations.core.composition.Transforms._find_dual_start_end
+--albumentations.augmentations.transforms.Normalize.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
+--albumentations.augmentations.transforms.Cutout.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
|      +--warnings.warn
+--albumentations.augmentations.transforms.RandomBrightnessContrast.__init__
+--albumentations.core.composition.Compose.__init__
|      +--albumentations.core.composition.BaseCompose.__init__
|      +--albumentations.augmentations.bbox_utils.BboxProcessor.__init__
|      |      +--albumentations.core.utils.DataProcessor.__init__
|      +--albumentations.core.composition.BboxParams.__init__
|      |      +--albumentations.core.utils.Params.__init__
|      +--albumentations.augmentations.keypoints_utils.KeypointsProcessor.__init__
|      |      +--albumentations.core.utils.DataProcessor.__init__
|      +--albumentations.core.composition.KeypointParams.__init__
|      |      +--albumentations.core.utils.Params.__init__
|      +--albumentations.core.composition.BaseCompose.add_targets
+--albumentations.augmentations.crops.transforms.RandomResizedCrop.__init__
|      +--albumentations.augmentations.crops.transforms._BaseRandomSizedCrop.__init__
|      |      +--albumentations.core.transforms_interface.BasicTransform.__init__
+--albumentations.augmentations.transforms.RandomBrightness.__init__
|      +--albumentations.augmentations.transforms.RandomBrightnessContrast.__init__
|      +--warnings.warn
+--albumentations.augmentations.transforms.MedianBlur.__init__
|      +--albumentations.augmentations.transforms.Blur.__init__
+--albumentations.imgaug.transforms.IAAAdditiveGaussianNoise.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
|      +--albumentations.core.transforms_interface.to_tuple
|      +--warnings.warn
+--albumentations.augmentations.geometric.resize.LongestMaxSize.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
+--albumentations.augmentations.transforms.RandomGamma.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
|      +--albumentations.core.transforms_interface.to_tuple
+--albumentations.augmentations.geometric.transforms.ShiftScaleRotate.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
|      +--albumentations.core.transforms_interface.to_tuple
+--albumentations.augmentations.transforms.Blur.__init__
+--albumentations.augmentations.transforms.VerticalFlip.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
+--albumentations.imgaug.transforms.IAAEmboss.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__
|      +--albumentations.core.transforms_interface.to_tuple
|      +--warnings.warn
+--albumentations.core.composition.BboxParams.__init__
+--albumentations.augmentations.transforms.HorizontalFlip.__init__
|      +--albumentations.core.transforms_interface.BasicTransform.__init__

Since all these functions have not been changed between any version for package "albumentations" from [1.0.1, 1.0.2, 1.0.3] and 1.0.0. Therefore, we believe it is safe to loosen the corresponding dependency.

No "timm-seresnet152d_320"

I was looking at the classification_aux folder and I was trying to call the get_encoder() function from the SMP library with the encoder_name = "timm-seresnet152d_320". However, I got an error saying that this wasn't part of the supported model encoders. The closest one was a "seresnet152". Below is the code and error I got. Could this be a version problem? Thanks!
se_resnet152d = get_encoder('timm-seresnet152d_320', in_channels=3, depth=5, weights='imagenet')

Regarding class weights in bce loss function

weight=torch.FloatTensor([0.2, 0.2, 0.3, 0.3]

Above parameter is corresponding to inputs in a batch.

To handle class imbalance, I think the parameter is pos_weight and not weight.

https://pytorch.org/docs/stable/generated/torch.nn.BCEWithLogitsLoss.html?highlight=bcewithlogitsloss#torch.nn.BCEWithLogitsLoss

Reproducing results from demo.ipynb

Is it possible for you to provide all the weights and checkpoints used for creating the demo.ipynb notebook for visualizing the model outputs? I was hoping to try the model on some data without having to train the model again.

dungnb1333 / siim-covid19-detection Goto Github PK

siim-covid19-detection's Introduction

SIIM-COVID19-Detection

1.INSTALLATION

2.DATASET

2.1 SIIM COVID 19 DATASET

2.2 EXTERNAL DATASET

3.SOLUTION SUMMARY

4.TRAIN MODEL

4.1 Classification

4.1.1 Multi head classification + segmentation

4.1.2 Lung Detector-YoloV5

4.2 Opacity Detection

4.2.1 YoloV5x6 768

4.2.2 EfficientDet D7 768

4.2.3 FasterRCNN FPN 768 & 1024

4.2.4 Ensembling + Pseudo labeling

4.2.5 Detection Performance

5.FINAL SUBMISSION

6.AWESOME RESOURCES

siim-covid19-detection's People

Contributors

Stargazers

Watchers

Forkers

siim-covid19-detection's Issues

Recommend Projects

Recommend Topics

Recommend Org