Publication

Abstract : Audio fingerprinting is a leap in music identification, copyright protection, and content-based retrieval. Existing methods rely primarily on spectral peaks and hash-based approaches, which are sensitive to noise, distortion, and temporal variations. In this paper, the introduction of a chaos-based audio fingerprinting system providing improved robustness and accuracy using chaotic transformations. The system derives spectral peaks from the Short-Time Fourier Transform (STFT) and applies a chaotic map transformation for feature hashing to ensure robust and unique fingerprint generation. Experimental results demonstrate that our method achieves 99.72% confidence in matching audio segments even in the presence of temporal variations. The proposed method demonstrates improved performance compared to existing methods in managing time shifts, Signal to Noise Ratio (SNR) and is a good candidate for real-world applications in music retrieval, audio forensics, and content recognition.