VGG16BNBackboneOptions
Options & Configuration · AiDotNet.ComputerVision.Detection.Backbones
Configuration options for the VGG16-BN detection backbone.
Inheritance: ModelOptions → NeuralNetworkOptions → ModelHyperparameterOptions → DetectionBackboneOptions → VGG16BNBackboneOptions
For Beginners
VGG16 with batch normalization is the feature extractor CRAFT’s text detector sits on. Its only knob is how many channels the input image has.
How It Works
The input channel count was a constructor parameter, so the backbone advertised no configuration surface and nothing could be set through an options object.
Constructors
| Constructor | Summary |
|---|---|
VGG16BNBackboneOptions() |
Methods
| Method | Summary |
|---|---|
void Validate() |
Throws when a value on this instance cannot produce a working backbone. |