Joint Dictionary Learning-Based Non-Negative Matrix Factorization for Voice Conversion to Improve Speech Intelligibility After Oral Surgery
Objective: This paper focuses on machine learning based voice conversion (VC) techniques for improving the speech intelligibility of surgical patients who have had parts of their articulators removed. Because of the removal of parts of the articulator, a patient's speech may be distorted and difficult to understand. To overcome this problem, VC methods can be applied to convert the distorted speech such that it is clear and more intelligible. To design an effective VC method, two key points must be considered: 1) the amount of training data may be limited (because speaking for a long time is usually difficult for postoperative patients); 2) rapid conversion is desirable (for better communication). Methods: We propose a novel joint dictionary learning based non-negative matrix factorization (JD-NMF) algorithm. Compared to conventional VC techniques, JD-NMF can perform VC efficiently and effectively with only a small amount of training data. Results: The experimental results demonstrate that the proposed JD-NMF method not only achieves notably higher short-time objective intelligibility (STOI) scores (a standardized objective intelligibility evaluation metric) than those obtained using the original unconverted speech but is also significantly more efficient and effective than a conventional exemplar-based NMF VC method. Conclusion: The proposed JD-NMF method may outperform the state-of-the-art exemplar-based NMF VC method in terms of STOI scores under the desired scenario. Significance: We confirmed the advantages of the proposed joint training criterion for the NMF-based VC. Moreover, we verified that the proposed JD-NMF can effectively improve the speech intelligibility scores of oral surgery patients..
| Medienart: |
Artikel |
|---|
| Erscheinungsjahr: |
2017 |
|---|---|
| Erschienen: |
2017 |
| Enthalten in: |
Zur Gesamtaufnahme - volume:64 |
|---|---|
| Enthalten in: |
IEEE transactions on biomedical engineering - 64(2017), 11, Seite 2584-2594 |
| Sprache: |
Englisch |
|---|
| Beteiligte Personen: |
Fu, Szu-Wei [VerfasserIn] |
|---|
| Links: |
|---|
| BKL: | |
|---|---|
| Themen: |
Data models |
| RVK: |
|---|
| doi: |
10.1109/TBME.2016.2644258 |
|---|
| funding: |
|
|---|---|
| Förderinstitution / Projekttitel: |
|
| PPN (Katalog-ID): |
OLC1997101998 |
|---|
| LEADER | 01000caa a2200265 4500 | ||
|---|---|---|---|
| 001 | OLC1997101998 | ||
| 003 | DE-627 | ||
| 005 | 20230513225537.0 | ||
| 007 | tu | ||
| 008 | 171125s2017 xx ||||| 00| ||eng c | ||
| 024 | 7 | |a 10.1109/TBME.2016.2644258 |2 doi | |
| 028 | 5 | 2 | |a PQ20171228 |
| 035 | |a (DE-627)OLC1997101998 | ||
| 035 | |a (DE-599)GBVOLC1997101998 | ||
| 035 | |a (PRQ)c1079-280beb99b9c61e11df2ae1d7f62031b90c77eac7353f969e598a2ae62c114130 | ||
| 035 | |a (KEY)0037705820170000064001102584jointdictionarylearningbasednonnegativematrixfacto | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
| 041 | |a eng | ||
| 082 | 0 | 4 | |a 620 |a 610 |q DE-600 |
| 084 | |a XA 48665 |q AVZ |2 rvk | ||
| 084 | |a 44.09 |2 bkl | ||
| 084 | |a 44.40 |2 bkl | ||
| 100 | 1 | |a Fu, Szu-Wei |e verfasserin |4 aut | |
| 245 | 1 | 0 | |a Joint Dictionary Learning-Based Non-Negative Matrix Factorization for Voice Conversion to Improve Speech Intelligibility After Oral Surgery |
| 264 | 1 | |c 2017 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a ohne Hilfsmittel zu benutzen |b n |2 rdamedia | ||
| 338 | |a Band |b nc |2 rdacarrier | ||
| 520 | |a Objective: This paper focuses on machine learning based voice conversion (VC) techniques for improving the speech intelligibility of surgical patients who have had parts of their articulators removed. Because of the removal of parts of the articulator, a patient's speech may be distorted and difficult to understand. To overcome this problem, VC methods can be applied to convert the distorted speech such that it is clear and more intelligible. To design an effective VC method, two key points must be considered: 1) the amount of training data may be limited (because speaking for a long time is usually difficult for postoperative patients); 2) rapid conversion is desirable (for better communication). Methods: We propose a novel joint dictionary learning based non-negative matrix factorization (JD-NMF) algorithm. Compared to conventional VC techniques, JD-NMF can perform VC efficiently and effectively with only a small amount of training data. Results: The experimental results demonstrate that the proposed JD-NMF method not only achieves notably higher short-time objective intelligibility (STOI) scores (a standardized objective intelligibility evaluation metric) than those obtained using the original unconverted speech but is also significantly more efficient and effective than a conventional exemplar-based NMF VC method. Conclusion: The proposed JD-NMF method may outperform the state-of-the-art exemplar-based NMF VC method in terms of STOI scores under the desired scenario. Significance: We confirmed the advantages of the proposed joint training criterion for the NMF-based VC. Moreover, we verified that the proposed JD-NMF can effectively improve the speech intelligibility scores of oral surgery patients. | ||
| 650 | 4 | |a Joint dictionary learning | |
| 650 | 4 | |a non-negative matrix factorization (NMF) | |
| 650 | 4 | |a Training | |
| 650 | 4 | |a Surgery | |
| 650 | 4 | |a Training data | |
| 650 | 4 | |a Dictionaries | |
| 650 | 4 | |a Speech | |
| 650 | 4 | |a sparse representation | |
| 650 | 4 | |a Data models | |
| 650 | 4 | |a Spectrogram | |
| 650 | 4 | |a voice conversion (VC) | |
| 700 | 1 | |a Li, Pei-Chun |4 oth | |
| 700 | 1 | |a Lai, Ying-Hui |4 oth | |
| 700 | 1 | |a Yang, Cheng-Chien |4 oth | |
| 700 | 1 | |a Hsieh, Li-Chun |4 oth | |
| 700 | 1 | |a Tsao, Yu |4 oth | |
| 773 | 0 | 8 | |i Enthalten in |t IEEE transactions on biomedical engineering |d New York, NY : IEEE, 1964 |g 64(2017), 11, Seite 2584-2594 |w (DE-627)129358452 |w (DE-600)160429-6 |w (DE-576)01473074X |x 0018-9294 |7 nnns |
| 773 | 1 | 8 | |g volume:64 |g year:2017 |g number:11 |g pages:2584-2594 |
| 856 | 4 | 1 | |u http://dx.doi.org/10.1109/TBME.2016.2644258 |3 Volltext |
| 856 | 4 | 2 | |u http://ieeexplore.ieee.org/document/7797132 |
| 912 | |a GBV_USEFLAG_A | ||
| 912 | |a SYSFLAG_A | ||
| 912 | |a GBV_OLC | ||
| 912 | |a SSG-OLC-TEC | ||
| 912 | |a SSG-OLC-PHA | ||
| 912 | |a SSG-OLC-DE-84 | ||
| 912 | |a SSG-OPC-PHA | ||
| 912 | |a GBV_ILN_70 | ||
| 912 | |a GBV_ILN_2061 | ||
| 912 | |a GBV_ILN_2410 | ||
| 936 | r | v | |a XA 48665 |
| 936 | b | k | |a 44.09 |q AVZ |
| 936 | b | k | |a 44.40 |q AVZ |
| 951 | |a AR | ||
| 952 | |d 64 |j 2017 |e 11 |h 2584-2594 | ||