| --- |
| license: apache-2.0 |
| tags: |
| - medical-imaging |
| - segmentation |
| - image-segmentation |
| - ct |
| - lung |
| - monai |
| - pytorch |
| library_name: pytorch |
| pipeline_tag: image-segmentation |
| --- |
| |
| # Lung ROI Segmentation — SegResNet (2D) |
|
|
| Per-slice 2D lung foreground (thoracic parenchyma) segmentation for |
| chest CT. Used as the ROI stage of a two-stage coarse-to-fine |
| pulmonary-nodule segmentation pipeline: its outputs are stacked into a |
| 3D lung bounding box that the downstream nodule model crops to. |
|
|
| ## Model details |
|
|
| - **Architecture**: [MONAI SegResNet](https://docs.monai.io/en/stable/networks.html#segresnet), 2D residual U-Net |
| - **Trainable parameters**: 6,904,081 |
| - **Input**: `(1, 256, 256)` axial CT slice, intensity-normalised to `[0, 1]` |
| - **Output**: `(1, 256, 256)` sigmoid map; foreground = lung tissue |
| - **Framework**: PyTorch + MONAI |
|
|
| ## Data |
|
|
| Trained on the `unified` split (patient-grouped, dataset-stratified) |
| of a unified corpus assembled from three public sources: |
|
|
| - NLST |
| - NSCLC-Radiomics |
| - LIDC-IDRI |
|
|
| Slices per split: **365 014 train / 64 097 val / — test held out**. |
|
|
| ## Validation metrics |
|
|
| Evaluated on 64 097 val slices (≈ 4.2 × 10⁹ pixels), micro-averaged at |
| 0.5 threshold on sigmoid output. |
|
|
| | Metric | Value | |
| |-------------|-------:| |
| | mIoU | 0.9803 | |
| | Accuracy | 0.9953 | |
| | Precision | 0.9797 | |
| | Recall | 0.9857 | |
| | Dice (F1) | 0.9827 | |
|
|
| ## How to load & run inference |
|
|
| ```python |
| import yaml, torch |
| from monai.networks.nets import SegResNet |
| |
| cfg = yaml.safe_load(open("config.yaml"))["model"] |
| model = SegResNet( |
| spatial_dims = cfg["spatial_dims"], |
| in_channels = cfg["in_channels"], |
| out_channels = cfg["out_channels"], |
| init_filters = cfg["init_filters"], |
| blocks_down = tuple(cfg["blocks_down"]), |
| blocks_up = tuple(cfg["blocks_up"]), |
| dropout_prob = cfg["dropout_prob"], |
| ) |
| state = torch.load("model.pth", map_location="cpu", weights_only=True) |
| model.load_state_dict(state) |
| model.eval() |
| |
| with torch.no_grad(): |
| x = torch.randn(1, 1, 256, 256) # (B, C, H, W) — replace with your CT slice |
| prob = torch.sigmoid(model(x)) |
| lung_mask = (prob > 0.5).to(torch.uint8) |
| ``` |
|
|
| ## Training recipe |
|
|
| - **Loss**: `DiceLoss(sigmoid=True, squared_pred=True)` |
| - **Optimizer**: Adam (lr = 1e-3, weight decay = 1e-5) |
| - **Scheduler**: CosineAnnealingLR (T_max = 100, η_min = 1e-6) |
| - **Batch size**: 16 |
| - **Samples/epoch**: 20 000 (random subsample of ~365 k train slices) |
| - **Epochs budget**: 100 | **best checkpoint at epoch 7 / 100** |
| - **Mixed precision**: bf16 |
| - **Seed**: 42 |
| - **Hardware**: 1 × NVIDIA H100 94 GB |
|
|
| Full config is included in this repo as `config.yaml`. |
|
|
| ## Reproducing training |
|
|
| Training code lives in an accompanying reproduction demo (published |
| separately). Once available, reproduce with: |
|
|
| ```bash |
| export DATA_ROOT=/path/to/unified # dir containing ct_2d/ and roi_sem_seg_2d/ |
| python train.py --config config.yaml |
| ``` |
|
|
| ## License & intended use |
|
|
| Model weights released under Apache 2.0. Training data was public but |
| covered by dataset-specific terms (NLST, NSCLC-Radiomics, LIDC-IDRI) — |
| users must comply with those separately when using the model on |
| comparable data. |
|
|
| **Not a medical device.** Not intended for clinical use. Research only. |
|
|
| ## Citation |
|
|
| Paper in preparation. |
|
|
