earticle

영어

Music Genre Recognition (MGR) plays a vital role in Music Information Retrieval (MIR), supporting content organization and recommendation in digital music platforms. This paper presents a concise survey of MGR, focusing on datasets, feature representations, and classification methods. Key feature types include timbral (e.g., MFCC), rhythmic, pitch-based, and spectrogram-based image features. Classification techniques span traditional machine learning (e.g., k-NN, SVM) and modern deep learning models (e.g., CNNs, RNNs, transformers). We summarize foundational, well-known, and highly cited works in the field, and compare widely used datasets. This survey aims to provide researchers with a structured understanding of the MGR domain, identify open challenges, and suggest future research directions.