Composer Identification in Classical Genre

Often it becomes very important to recognize the music composer from the composition by just listening to it. Here in this paper we try to create a robust architecture to address the issue of composer identification. Along with the solutions, we have prepared a dataset for the same. We have extracted multiple numbers of music of each of the composers namely Chopin, Beethoven and Mozart. The dataset has been created by taking thirty seconds of each of the audio tracks of each composer. The performance evaluation is done for different machine learning model, 1-D CNN model and different state of the art transfer learning models with fine tuning based on average accuracy, recall, precision and F1 measure.