Polaris improves audio fingerprinting without training or large databases
POLARIS: Training-Free Audio Fingerprinting with Saliency-Based Landmarks and Delaunay Grouping
SoundComputer Vision and Pattern Recognition
Summary
Recognizing songs even when the sound quality changes is hard. The authors present Polaris, a system that finds special audio features without needing to learn from examples. It groups these features cleverly to match songs accurately from noisy or altered recordings. Polaris works better than other methods without training and even beats some neural approaches on real-world music recordings.
What this means in practice
- •For music app developers: Build music recognition features that can identify songs from distorted or re-recorded audio without large training datasets.$Commercial implications: Enables commercial music ID apps to deliver accurate recognition using lightweight and training-free methods, reducing development and operational costs.
- •For audio archive managers: Organize and catalog large music collections by matching recordings even with quality variations using an efficient indexing and matching system.
Authors
Jiheng Li
Abstract
This work presents POLARIS, a training-free audio fingerprinting system that selects landmarks from a locally normalized saliency field and groups them into sparse fingerprints using Delaunay triangulation. To deal with query distortion, POLARIS adds fingerprints from two-hop Delaunay neighborhoods only at query time, without enlarging the reference index. An adaptive configuration applies this expansion only when the original fingerprints do not produce a confident match. We evaluate POLARIS on synthetic distortions from the public PEX Hard Medium benchmark, excluding queries with pitch or tempo shifts, and on a new benchmark of real re-recorded music. POLARIS achieves the best performance among the evaluated training-free methods on both benchmarks. On the real recordings, its adaptive configuration also outperforms the neural NMFP baseline with a comparable measured query time and a smaller logical reference payload. Code, dataset, and instructions for reproducing all experiments are available at https://github.com/JihengLi/POLARIS.git.