Lung Nodule Segmentation — SegResNet (3D, WIDE)
Voxel-level segmentation of pulmonary nodules in 3D chest CT volumes,
cropped to the lung region. Wide-capacity SegResNet
(init_filters = 32, ~83 M params) retrained on the unified
split.
This is the "wide SegResNet" architecture from the paper, directly
comparable to the other two-stage nodule models (segresnet_small,
dynunet) — all of which share the same train / val / test split.
Peaked at val Dice 0.5711 (per-case mean) at epoch 489 / 1000.
Model details
- Architecture: MONAI SegResNet, 3D residual U-Net
- Trainable parameters: 82,637,282
- Input:
(1, 256, 256, 256) CT crop, intensity-normalised to [0, 1], resampled from a per-series lung bbox (padding = 20 vox)
- Output:
(2, 256, 256, 256) softmax logits — class 0 = background, class 1 = nodule
- Framework: PyTorch + MONAI
Data
Trained on the unified split (patient-grouped, dataset-stratified):
| Source | Role |
|---|
| NLST | train + val + test |
| NSCLC-Radiomics | train + val + test |
| LIDC-IDRI | train + val + test |
Split sizes: 1 683 train / 297 val / 325 test (held out).
Note on bboxes. This model was originally trained with a regenerated
bbox JSON (tighter lung crops) that differs from the
bboxes_unified.json bundled with the small and DynUNet variants.
Reproducing from the bundled bboxes gives small variance from the
released checkpoint but does not materially affect training.
Validation & test metrics
Test split (325 series, ~5.5 × 10⁹ voxels, micro-averaged):
| Metric | Value |
|---|
| mIoU | 0.7482 |
| Accuracy | 0.9996 |
| Precision | 0.5766 |
| Recall | 0.7818 |
| Dice / F1 (micro) | 0.6637 |
Per-case Dice distribution: mean 0.5706, median 0.6304, p25 0.4069,
p75 0.7834.
Comparison to the other two-stage nodule models on the same test set
| Model | mIoU | Precision | Recall | Dice (micro) | Per-case Dice mean |
|---|
nodule-segresnet-3d-small | 0.7391 | 0.539 | 0.812 | 0.6478 | 0.5240 |
nodule-dynunet-3d | 0.7594 | 0.624 | 0.757 | 0.6838 | 0.5560 |
| this model (wide) | 0.7482 | 0.577 | 0.782 | 0.6637 | 0.5706 |
On voxel-level metrics (mIoU / Dice_micro), DynUNet wins; on
per-case Dice mean (equal weight per patient regardless of nodule size),
this wide-SegResNet variant wins. Which is "best" depends on the
metric: large-nodule cases dominate the micro-averaged numbers, while
per-case-averaged metrics weigh each patient equally.
How to load & run inference
1import yaml, torch
2from monai.networks.nets import SegResNet
3
4cfg = yaml.safe_load(open("config.yaml"))["model"]
5model = SegResNet(
6 spatial_dims = cfg["spatial_dims"],
7 in_channels = cfg["in_channels"],
8 out_channels = cfg["out_channels"], # 2
9 init_filters = cfg["init_filters"], # 32 (wide)
10 blocks_down = tuple(cfg["blocks_down"]),
11 blocks_up = tuple(cfg["blocks_up"]),
12 dropout_prob = cfg["dropout_prob"],
13)
14state = torch.load("model.pth", map_location="cpu", weights_only=True)
15model.load_state_dict(state)
16model.eval()
17
18with torch.no_grad():
19 # Input: (B, 1, 256, 256, 256), lung-bbox-cropped CT resampled to 256³
20 x = torch.randn(1, 1, 256, 256, 256)
21 logits = model(x) # (B, 2, D, H, W)
22 pred_class = logits.argmax(dim=1) # (B, D, H, W) in {0, 1}
23 nodule_mask = (pred_class == 1).to(torch.uint8)
This model expects
lung-bbox-cropped input. A separate 2D ROI model
is needed to produce that bbox — see the accompanying ROI checkpoints
(
szabopeter/roi-segresnet-2d
or
szabopeter/roi-swinunetr-2d),
or use the demo's
inference.py which chains them for you.
Training recipe
- Loss: Focal Tversky + weighted CE (α=0.3, β=0.7, γ=2.0, λ_CE=0.3,
ce_nodule_weight = 100)
- Optimizer: Adam (lr=1e-5, wd=1e-5)
- Scheduler: CosineAnnealingLR (T_max=1000, η_min=1e-6)
- Batch size: 2
- Epochs: 1000 | best checkpoint at epoch 489 / 1000
- Mixed precision: bf16
- Seed: 42
- Hardware: 1 × NVIDIA H100 94 GB
- Wall-clock: ≈ 12 days
Full config is included in this repo as config.yaml.
Companion models
szabopeter/nodule-segresnet-3d-small — small SegResNet (init_filters=16, ~20 M params)szabopeter/nodule-dynunet-3d — DynUNet / 3D U-Net (~31 M params, best voxel-level metrics)
License & intended use
Apache 2.0. Training data was public but subject to dataset-specific
terms (NLST, NSCLC-Radiomics, LIDC-IDRI).
Not a medical device. Research use only.
Citation
Paper in preparation.
