Recognition and Classification of Accompanying Audios of Kathak Dance

In our work a novel method has been implemented to recognize and classify the accompanying audios used in Indian Classical Dance(ICD) variation of Kathak. To achieve this Kathak Ladi footwork audio dataset have been used. This audio dataset has been analyzed and feature extraction is performed for the classification task. Hence audio signal and speech processing plays an important roll here. Beat detection is carried out to detect onset of the beats and then beat tracking is done on the audio signals. Then recognition and classification of Ladi audios is done based on rhythms and mfccs into 7 classes namely L2, L3, L4, L5, L6, L7 and L8 using machine learning based approach.