Efficient Cepstral Normalization for Robust Speech Recognition
By Fu-hua Liu, Richard M. Stern, Xuedong Huang, A. Acero. 1993
In this paper we describe and compare the performance of a series of cepstrum-based procedures that enable the CMU SPHINX-II speech recognition system to maintain a high level of recognition accuracy over a wide variety of acoustical environments. We describe the MFCDCN algorithm, an environment-independent extension of the efficient SDCN and FCDCN algorithms developed previously. We compare the performance of these algorithms with the very simple RASTA and cepstral mean normalization procedures, describing the performance of these algorithms in the context of the 1992 DARPA CSR evaluation using secondary microphones, and in the DARPA stress-test evaluation.
Subfields: Natural Language Processing
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Natural Language Processing
- Thomas G. Dietterich
- Gerasimos Potamianos
- Ashutosh Garg
- Atri Rudra
- Christopher R'e
- Chalapathy Neti
- Tri Dao
- Andrew W. Senior
Citations in the corpus
- Benchmarking Neural Network Robustness to Common Corruptions and Perturbations (2019, methodology): “Additional robustness has been achieved through pre-processing techniques such as standardizing the statistics of the input (Liu et al., 1993; Torre et al., 2005; Harvilla & Stern, 2012; Kim & Stern, 2016).”
- Benchmarking Neural Network Robustness to Common Corruptions and Surface Variations (2018, methodology): “Additional robustness has been achieved through pre-processing techniques such as standardizing the statistics of the input [40, 54, 18, 35].”
- Recent advances in the automatic recognition of audiovisual speech (2003, background): “Variations between speakers and recording conditions can be somewhat remedied by visual feature mean normalization (FMN), i.e., by subtraction of the vector mean over each utterance [ 3 ], [79].”
- Kaleidoscope: An Efficient, Learnable Representation For All Structured Linear Maps (2020, methodology): “, 1937)) by multiplication with a specific banded matrix of dimension 512× 23, and the entrywise logarithm of the output is taken (the 23 outputs are called the filters), and (vii) cepstral mean and variance normalization (Liu et al., 1993) is applied.”
- Audio-Visual Automatic Speech Recognition: An Overview (2004, unknown): “In a traditional audio front end, cepstral mean subtraction is often employed to enhance robustness to speaker and environment variations (Liu et al., 1993; Young et al., 1999).”