| --- |
| language: en |
| license: mit |
| tags: |
| - medical-imaging |
| - segmentation |
| - in-context-learning |
| - interactive-segmentation |
| - tensorflow |
| - keras |
| - ct |
| - mri |
| - onnx |
| library_name: tensorflow |
| pipeline_tag: image-segmentation |
| --- |
| |
| # LISP-Net — Lightweight In-Context Slice Propagator Network |
|
|
| LISP-Net is a lightweight, purely convolutional framework for interactive volumetric medical image segmentation. Instead of sparse clicks, it uses a single dense 2D prompt — a reference image paired with a full mask — to derive structural guidance directly from the individual patient. |
|
|
| ## Model Details |
|
|
| - **Architecture:** Asymmetrical dual-encoder U-Net with multi-resolution SE channel-attention |
| - **Parameters:** ~28M |
| - **Input:** Query image (128×128) + Prompt (reference image + binary mask, stacked as 2 channels) |
| - **Output:** Binary segmentation probability map (128×128) |
| - **Training data:** 208 patients across 7 datasets (NAKO, TotalSegmentator, MSD, BraTS-GLI, BraTS-MEN-RT, TopCoW MR, TopCoW CT) |
|
|
| ## Performance |
|
|
| | Benchmark | vs. | Result | |
| |-----------|-----|--------| |
| | 2D (offset ±5) | UniverSeg | 0.798 vs. 0.597 DSC | |
| | 2D (offset ±12) | UniverSeg | 0.704 vs. 0.569 DSC | |
| | 3D (SSF only) | nnInteractive | 0.665 vs. 0.705 Vol. DSC | |
| | 3D (interactive) | nnInteractive | 0.879 vs. 0.810 Vol. DSC | |
|
|
| Peak GPU memory: 164–362 MB. Per-slice latency: ~14 ms (GPU) / ~150 ms (CPU). |
|
|
| ## Usage |
|
|
| ### Python (Keras) |
|
|
| ```python |
| from inference.predictor import PromptUNetPredictor |
| |
| # Downloads from Hugging Face automatically on first use |
| predictor = PromptUNetPredictor("Machauer-P/lisp-net") |
| |
| mask = predictor.predict(query_image, prompt) |
| ``` |
|
|
| ### ONNX (Browser / ONNX Runtime) |
|
|
| Download `lisp_net_332.onnx` and use with ONNX Runtime or integrate into a web application. |
|
|
| ## Intended Use |
|
|
| - Interactive volumetric segmentation of CT and MRI |
| - Zero-shot generalization to novel anatomical targets without retraining |
| - Clinical research and annotation workflows |
|
|
| ## Limitations |
|
|
| - Requires a full 2D dense annotation as initial prompt (higher upfront effort than sparse clicks) |
| - Optimized for medium-range propagation (offset ≤16 slices); distant slices may need prompt refreshes |
|
|
| ## License |
|
|
| MIT |
|
|
| ## Citation |
|
|
| Paper forthcoming. See [Machauer-P/lisp-net](https://github.com/Machauer-P/lisp-net) for updates. |
|
|
