DNN-Based Geometry-Invariant DOA Estimation With Microphone Positional Encoding and Complexity Gradual Training
Min-Sang Baek, Joon‐Hyuk Chang, Israel Cohen · IEEE Transactions on Audio Speech and Language Processing · 2025
Recent deep neural network (DNN)-based direction-of-arrival (DOA) estimation methods demonstrate greater robustness compared to conventional methods. However, most DNNs are designed for specific microphone arrays, requiring retraining for different geometries. Although some geometry-invariant methods employ conventional features, they often incur high computational costs and are prone to interference. This paper proposes ageometry-invariant DOA estimation network(GI-DOAEnet). It employsmicrophone positional encodings(MPEs) that modulate microphone spherical coordinates using sinusoidal functions to provide unique geometric information. Combining MPEs and channel-wise latent features, the network captures spatio-temporal correlations through geometry-invariant modules, ultimately producing spatial spectra. To train GI-DOAEnet effectively with diverse geometries, acomplexity gradual trainingstrategy is introduced, integratingdeeply supervised curriculum learningwith a novelmulti-stage geometry learningmethod. This gradually increases task difficulty by training through varying soft labels and staged transitions from fixed to dynamic geometries. GI-DOAEnet achieves superior performance over baselines in terms of degree error and accuracy across diverse acoustic environments, while reducing FLOPS and inference time by eliminating pair-wise features and employing channel-wise aggregation.