Multidimensional acoustic feature fusion with attention-guided light weighted CNN for improved speaker recognition

Key Points

• Enhanced speaker recognition achieved using an attention mechanism to process acoustic features.
• Key improvement demonstrated by experimental results showing an x% increase in accuracy over baseline methods.
• Utilizing a convolutional neural network (CNN), the model effectively fuses multiple acoustic features for optimal recognition.
• The findings support further exploration of lightweight models in real-time speaker identification applications.