Convolutional Neural Network and Tawny Flamingo Optimization with Cubic Chaos Strategy for Classification of Music Genre

Music genre designations are valuable for organizing songs, albums, and artists with similar musical traits into broader categories, with one of the most effective applications being the enhancement of music recommendations and consumer accessibility. However, traditional methods struggle with accurate genre classification, impacting classification performance. This paper proposes Convolutional Neural Networks (CNN) for genre classification, leveraging its ability to model complex relationships in extracted features, while the Tawny Flamingo Optimization with Cubic Chaos strategy (CC-TFO) efficiently fine-tuned data, achieving improved accuracy. The data is acquired from the GTZAN dataset and pre-processed using normalization to uniformly analyze the music signal. Feature extraction is carried out using Mel-Frequency Cepstral Coefficients (MFCC), Root Mean Square Energy (RMS), and Mean Curve Length to capture spectral and audio signals and evaluate the energy level in dynamic range. Further, CNN is employed for precise feature learning, while TFO is used to optimize parameters, with the CC strategy accelerating tuning process for ensuring better classification accuracy. The proposed method achieves an accuracy of 99.20% on the GTZAN dataset, outperforming the existing Recurrent Neural Network (RNN) technique.