Partial AUC Optimization Based Deep Speaker Embeddings With Class-center Learning For Text-independent Speaker Verification

Abstract

Deep embedding based text-independent speaker verification has demonstrated superior performance to traditional methods in many challenging scenarios. Its loss functions can be generally categorized into two classes, i.e., verification and identification. The verification loss functions match the pipeline of speaker verification, but their implementations are difficult. Thus, most state-of-the-art deep embedding methods use the identification loss functions with softmax output units or their variants. In this paper, we propose a verification loss function, named the maximization of partial area under the Receiver-operating-characteristic (ROC) curve (pAUC), for deep embedding based text-independent speaker verification. We also propose a class-center based training trial construction method to improve the training efficiency, which is critical for the proposed loss function to be comparable to the identification loss in performance. Experiments on the Speaker in the Wild (SITW) and NIS

Stats

• citations18
• S2 citations—
• github stars0
• HF likes0
• heat score9.59
• arxiv keybai2019partial

Related papers

• Optimization Of The Area Under The ROC Curve Using Neural Network Supervectors For Text-dependent Speaker Verification (2019)10.21
• Speaker Verification By Partial AUC Optimization With Mahalanobis Distance Metric Learning (2019)9.92
• Feature Enhancement With Deep Feature Losses For Speaker Verification (2019)10.61
• Unified Hypersphere Embedding For Speaker Recognition (2018)0.00
• Large Margin Softmax Loss For Speaker Verification (2019)14.66
• Angular Softmax Loss For End-to-end Speaker Verification (2018)11.19
• On Deep Speaker Embeddings For Text-independent Speaker Recognition (2018)11.93
• Centroid-based Deep Metric Learning For Speaker Recognition (2019)13.79