BiomedParse<T>
Models & Types · AiDotNet.ComputerVision.Segmentation.Medical
BiomedParse: Biomedical image parsing with text prompts.
Inheritance: NeuralNetworkBase<T> → SegmentationModelBase<T> → MedicalSegmentationBase<T> → BiomedParse<T>
For Beginners
Text-prompted biomedical image segmentation. Multi-modality biomedical parsing.
Common use cases:
- Text-prompted biomedical image segmentation
- Multi-modality biomedical parsing
- Detection and recognition in biomedical images
- Joint segmentation-detection-recognition
How It Works
Technical Details:
- Text-prompted segmentation for biomedical images
- Joint segmentation, detection, and recognition in one model
- Trained on 6M+ triples across 9 imaging modalities
- GPT-4 assisted harmonization of biomedical datasets
Reference: Zhao et al., “BiomedParse: a biomedical foundation model for image parsing of everything everywhere all at once”, Nature Methods 2024.
Example
// Create a BiomedParse model for text-prompted biomedical segmentation
var architecture = new NeuralNetworkArchitecture<double>(
inputType: InputType.ThreeDimensional,
taskType: NeuralNetworkTaskType.BinaryClassification,
inputHeight: 1024, inputWidth: 1024, inputDepth: 3, outputSize: 1);
var model = new BiomedParse<double>(architecture);
// Every tunable is now set through the options object; these are the defaults:
var custom = new BiomedParse<double>(architecture,
options: new BiomedParseOptions { NumClasses = 1, DropRate = 0.1 });
// Or load a pre-trained ONNX model for multi-modality biomedical parsing
var onnxModel = new BiomedParse<double>(architecture, "biomedparse.onnx");
Constructors
| Constructor | Summary |
|---|---|
BiomedParse(NeuralNetworkArchitecture<T> architecture, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer, ILossFunction<T> lossFunction, BiomedParseOptions options) |
Initializes BiomedParse in native (trainable) mode. |
BiomedParse(NeuralNetworkArchitecture<T> architecture, string onnxModelPath, BiomedParseOptions options) |
Initializes BiomedParse in ONNX (inference-only) mode. |
Properties
| Property | Summary |
|---|---|
bool Supports3D { get; } |
BiomedParse is a 2D slice model; it does not process volumes. |
Methods
| Method | Summary |
|---|---|
MedicalSegmentationResult<T> SegmentFewShot(Tensor<T> queryImage, Tensor<T> supportImages, Tensor<T> supportMasks) |
Not supported: BiomedParse has no few-shot pathway. |
MedicalSegmentationResult<T> SegmentSlice(Tensor<T> slice) |
|
MedicalSegmentationResult<T> SegmentVolume(Tensor<T> volume) |
|
ModelMetadata<T> GetModelMetadata() |
Collects metadata describing this model’s configuration. |