Deep Speaker Feature Learning for Text-Independent Speaker Verification
By Lantian Li, Yixiang Chen, Ying Shi, Zhiyuan Tang, Dong Wang. 2017
89 citations recorded by the source corpus.
Recently deep neural networks (DNNs) have been used to learn speaker features. However, the quality of the learned features is not sufficiently good, so a complex back-end model, either neural or probabilistic, has to be used to address the residual uncertainty when applied to speaker verification, just as with raw features. This paper presents a convolutional time-delay deep neural network structure (CT-DNN) for speaker feature learning. Our experimental results on the Fisher database demonstrated that this CT-DNN can produce high-quality speaker features: even with a single feature (0.3 seconds including the context), the EER can be as low as 7.68%. This effectively confirmed that the speaker trait is largely a deterministic short-time property rather than a long-time distributional pattern, and therefore can be extracted from just dozens of frames.
Subfields: Artificial Intelligence, Signal Processing
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Yoshua Bengio
- Sergey Levine
- Quoc V. Le
- Mohit Bansal
- Ruslan Salakhutdinov
- Pieter Abbeel
- Kyunghyun Cho
- Jianfeng Gao
In Signal Processing
- Aäron van den Oord
- Junichi Yamagishi
- Oriol Vinyals
- Heiga Zen
- Yoshua Bengio
- Dong Wang
- Lior Wolf
- Yannis Assael
Citations in the corpus
- Sample Efficient Adaptive Text-to-Speech (2018, background): “Meta-learning has a long history (Harlow, 1949; Thrun and Pratt, 2012), and recent studies include efforts to learn optimization processes (Andrychowicz et al., 2016; Chen et al., 2017) that have been shown to extend naturally to the few-shot setting (Ravi and Larochelle, 2016).”
- Speaker recognition based on deep learning: An overview (2020, unknown): “F-bank TDNN [14, 65, 66, 67]; ResNet [68, 69, 70, 71]; VGGNet [72]; Inception-resnet-v1 [63, 73, 74]; Others [75, 76].”
- CN-Celeb: A Challenging Chinese Speaker Recognition Dataset (2019, background): “Resorting to the power of deep neural networks (DNNs) in representation learning, these methods can remove unwanted uncertainties by propagating speech signals through the DNN layer by layer and retain speaker-relevant features only [10].”
- Unspeech: Unsupervised Speech Context Embeddings (2018, background): “Recently, supervised neural network based speaker embeddings also showed good speaker discriminative properties [6, 7, 8], particularly on short utterances.”