Computer Science · Signal Processing
Aural Signatures: Audio Fingerprinting Techniques for Real-Time Audio Recognition
B. Sharan Nripesh Reddy, B. Sai Venkat, Irigineni Manohar, S Abhishek, T Anjali
See on the mapAbstract
Audio fingerprinting is an essential technique for copyright protection and music recognition. The capacity to accurately categorize audio content and spot possible copyright violations is vital for musicians, content creators, and streaming services. With its valuable solutions for copyright protection, content monitoring, and music recommendation, this study has significant implications for the music industry. In the research, the research focused on developing robust audio fingerprinting methods for precise audio content recognition. These tactics have shown to be remarkably accurate and quite effective. The approach is based on the use of strong hashing techniques and the extraction of various audio aspects. Long Short-Term Memory (LSTM) along with SVM, KNN, and CNN were employed in the research. Notably, LSTM outperformed other models, showcasing an outstanding accuracy rate of 93 %. The results hold particular promise for massive music platforms and instantaneous content matching. The proposed methods have demonstrated the effectiveness of these strategies in identifying audio material, even under challenging conditions.