PeakNetFP<T>

Models & Types · AiDotNet.Audio.Fingerprinting

PeakNetFP spectral peak-based neural audio fingerprinting model.

Inheritance: NeuralNetworkBase<T> → AudioNeuralNetworkBase<T> → PeakNetFP<T>

Implements: IAudioFingerprinter<T>

For Beginners

PeakNetFP identifies songs by finding the “peaks” in their sound spectrum (like the loudest frequencies at each moment) and then uses AI to turn those peaks into a compact code. It’s a hybrid approach combining classical peak picking with neural encoding.

Usage:

How It Works

PeakNetFP combines traditional spectral peak picking with a neural network for robust audio fingerprinting. It detects spectral peaks in the spectrogram and uses a CNN to encode peak constellations into compact binary hashes, offering both speed and robustness.

Constructors

Constructor Summary
PeakNetFP(NeuralNetworkArchitecture<T> architecture, PeakNetFPOptions options, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer) Creates a PeakNetFP model in native training mode.
PeakNetFP(NeuralNetworkArchitecture<T> architecture, string modelPath, PeakNetFPOptions options) Creates a PeakNetFP model in native training mode.

Properties

Property Summary
int FingerprintLength { get; }
string Name { get; }

Methods

Method Summary
AudioFingerprint<T> Fingerprint(Tensor<T> audio)
AudioFingerprint<T> Fingerprint(Vector<T> audio)
IReadOnlyList<FingerprintMatch> FindMatches(AudioFingerprint<T> query, AudioFingerprint<T> reference, int minMatchLength)
ModelMetadata<T> GetModelMetadata()
double ComputeSimilarity(AudioFingerprint<T> fp1, AudioFingerprint<T> fp2)