SceneClassifierOptions

Options & Configuration ยท AiDotNet.Audio.Classification

Options for acoustic scene classification with the DCASE 2017 baseline system (Mesaros et al. 2017, section 2.2): 40 log mel-band energies over 40 ms frames with 50% overlap, a 5-frame context, and a two-layer MLP of 50 units with 20% dropout.

Inheritance: ModelOptions โ†’ SceneClassifierOptions

For Beginners

These options configure the SceneClassifier model. Default values follow the DCASE 2017 baseline system.

Constructors

Constructor Summary
SceneClassifierOptions() Initializes a new instance with default values.
SceneClassifierOptions(SceneClassifierOptions other) Initializes a new instance by copying from another instance.

Properties

Property Summary
OnnxModelOptions OnnxOptions { get; set; } ONNX model options.
double DropoutRate { get; set; } Dropout after each hidden layer.
int ContextFrames { get; set; } Frames concatenated into one feature vector, centred on the frame classified.
int FftSize { get; set; } Analysis frame length in samples.
int HiddenUnits { get; set; } Units of each hidden dense layer.
int HopLength { get; set; } Hop length in samples.
int NumHiddenLayers { get; set; } Hidden dense layers.
int NumMels { get; set; } Number of mel bands.
int NumMfccs { get; set; } Number of MFCCs in the descriptive SceneFeatures of a result (not a classifier input).
int SampleRate { get; set; } Audio sample rate.
int TopK { get; set; } Number of top predictions to return.
string ModelPath { get; set; } Path to ONNX model file (optional).
string[] CustomScenes { get; set; } Custom scene labels (optional).