TextureClassificationModel
Contains all the information necessary to run a trained machine learning model for performing a texture-based segmentation.
The object publicly exposes only the most relevant information of the model. Other important information is stored in this model, such as the classification weights, but is not accessible.
The object publicly exposes only the most relevant information of the model. Other important information is stored in this model, such as the classification weights, but is not accessible.
Object members
| Measurement name | Description | Element type | Indexing |
|---|---|---|---|
| computeMode | The type of dimensional image to process with the model (2D or 3D). | String | None |
| coocRadius | The radius, in pixels, of the circular neighborhood used by the co-occurrence features. | Unsigned integer | None |
| coocTextonShape | The shape of the co-occurrence texton; that is, the pattern defined by the set of co-occurrence vectors. | String | None |
| listOfRadius | The list of radius, in pixels, of the circular neighborhoods used for computing textural features. | Unsigned integer | [radiusIndex] |
| numOfClasses | The number of classes learned by the model. | Unsigned integer | None |
| selectedFeature | The list of features retained by the training to discrimate the textures. | String | [featureIndex] |
| separationPower | The parameter controlling the rejection criteria of the feature selection algorithm. | Floating point | None |
| featureGroup | The feature groups selected when initializing the model. | String | [radiusIndex] |