Feature fusion based automatic chord recognition model: BTC-FDAA-FGF

Abstract

Automatic chord recognition is a significant topic in the field of Music Information Retrieval (MIR). This paper introduces a novel feature fusion method combining Hybrid Constant-Q Transform (HCQT) and adaptive attention for chord detection, especially with a focus on improving the accuracy of chord detection of rare chord classes. Serving as one of the cornerstone features of music, the chords obtained by chord recognition algorithms are the basis of many high-level semantic tasks. At present, a severe class imbalance problem exists in the domain of automatic chord recognition. The recognition accuracy of rare chords is much lower than that of common chords, which significantly affect the overall performance of chord recognition algorithms. In this paper, a chord recognition algorithm based on feature fusion is proposed. First, in the feature extraction part, Hybrid Constant-Q Transform (HCQT) is introduced to assist with Constant-Q Transform(CQT) to obtain richer and finer musical signal features, enabling better tracking of overtones. Next, in the chord estimation part, the frequency-domain adaptive attention (FDAA) mechanism is used to enhance feature saliency, ensuring that the network can adaptively adjust the weights for different frequency components when training. Thereby frequency-domain features that contain important information can be selectively enhanced. The enhanced features are then fed into an aggregation module that integrates a bidirectional self-attention module and Fourier transform module, enabling more effective capture of fine-grained features, global context information, and periodic structures in chords. The experimental result shows that proposed algorithm outperforms existing mainstream baseline methods by 1.2% to 2.2% on the MIREX metrics, validating the effectiveness of the algorithm.