🎙️ 3D Few-Shot Speaker Enrolment: Embedding Averaging, Cosine Scoring and Equal Error Rate
A 3D view of few-shot speaker enrolment: noisy recordings scatter around a speaker, their average sharpens the template, and target and impostor cosine scores form two bell-shaped towers split by a threshold; closed-form means, spreads and equal error rate are checked against Monte Carlo (synthetic embeddings).
⚙ Under the hood
A 3D view of few-shot speaker enrolment: noisy recordings scatter around a speaker, their average sharpens the template, and target and impostor cosine scores form two bell-shaped towers split by a threshold; closed-form means, spreads and equal error rate are checked against Monte Carlo (synthetic embeddings).
Speaker VerificationSpeaker EmbeddingCosine SimilarityEqual Error RateFew-Shot LearningNoise AveragingHigh-Dimensional GeometryVoice Biometrics
3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install