Adaptive noise dictionary design for noise robust exemplar matching of speech. Yilmaz, E., Van hamme, H., & Gemmeke, J. F. In 2015 23rd European Signal Processing Conference (EUSIPCO), pages 1681-1685, Aug, 2015.
Paper doi abstract bibtex This paper investigates an adaptive noise dictionary design approach to achieve an effective and computationally feasible noise modeling for the noise robust exemplar matching (N-REM) framework. N-REM approximates noisy speech segments as a linear combination of multiple length exemplars in a sparse representation (SR) formulation. Compared to the previous SR techniques with a single overcomplete dictionary, N-REM uses smaller dictionaries containing considerably fewer noise exemplars. Hence, the noise exemplars have to be selected with care to accurately model the spectrotem-poral content of the actual noise conditions. For this purpose, in a previous work, we introduced a noise exemplar selection stage before performing recognition which extracts noise exemplars from a few noise-only training sequences chosen for each target noisy utterance. In this work, we explore the impact of the several design parameters on the recognition accuracy by evaluating the system performance on the CHIME-2 and AURORA-2 databases.
@InProceedings{7362670,
author = {E. Yilmaz and H. {Van hamme} and J. F. Gemmeke},
booktitle = {2015 23rd European Signal Processing Conference (EUSIPCO)},
title = {Adaptive noise dictionary design for noise robust exemplar matching of speech},
year = {2015},
pages = {1681-1685},
abstract = {This paper investigates an adaptive noise dictionary design approach to achieve an effective and computationally feasible noise modeling for the noise robust exemplar matching (N-REM) framework. N-REM approximates noisy speech segments as a linear combination of multiple length exemplars in a sparse representation (SR) formulation. Compared to the previous SR techniques with a single overcomplete dictionary, N-REM uses smaller dictionaries containing considerably fewer noise exemplars. Hence, the noise exemplars have to be selected with care to accurately model the spectrotem-poral content of the actual noise conditions. For this purpose, in a previous work, we introduced a noise exemplar selection stage before performing recognition which extracts noise exemplars from a few noise-only training sequences chosen for each target noisy utterance. In this work, we explore the impact of the several design parameters on the recognition accuracy by evaluating the system performance on the CHIME-2 and AURORA-2 databases.},
keywords = {acoustic noise;speech recognition;adaptive noise dictionary design;feasible noise modeling;noise robust exemplar matching framework;multiple length exemplars linear combination;SR techniques;sparse representation formulation;noise-only training sequences;target noisy utterance;recognition;N-REM;noisy speech segments;Dictionaries;Speech;Training;Noise measurement;Hidden Markov models;Signal to noise ratio;Adaptation models;template matching;noise-robustness;automatic speech recognition;sparse representations;exemplar selection},
doi = {10.1109/EUSIPCO.2015.7362670},
issn = {2076-1465},
month = {Aug},
url = {https://www.eurasip.org/proceedings/eusipco/eusipco2015/papers/1570102603.pdf},
}
Downloads: 0
{"_id":"AQCb8iW5anLo28RMT","bibbaseid":"yilmaz-vanhamme-gemmeke-adaptivenoisedictionarydesignfornoiserobustexemplarmatchingofspeech-2015","authorIDs":[],"author_short":["Yilmaz, E.","Van hamme, H.","Gemmeke, J. F."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["E."],"propositions":[],"lastnames":["Yilmaz"],"suffixes":[]},{"firstnames":["H."],"propositions":[],"lastnames":["Van hamme"],"suffixes":[]},{"firstnames":["J.","F."],"propositions":[],"lastnames":["Gemmeke"],"suffixes":[]}],"booktitle":"2015 23rd European Signal Processing Conference (EUSIPCO)","title":"Adaptive noise dictionary design for noise robust exemplar matching of speech","year":"2015","pages":"1681-1685","abstract":"This paper investigates an adaptive noise dictionary design approach to achieve an effective and computationally feasible noise modeling for the noise robust exemplar matching (N-REM) framework. N-REM approximates noisy speech segments as a linear combination of multiple length exemplars in a sparse representation (SR) formulation. Compared to the previous SR techniques with a single overcomplete dictionary, N-REM uses smaller dictionaries containing considerably fewer noise exemplars. Hence, the noise exemplars have to be selected with care to accurately model the spectrotem-poral content of the actual noise conditions. For this purpose, in a previous work, we introduced a noise exemplar selection stage before performing recognition which extracts noise exemplars from a few noise-only training sequences chosen for each target noisy utterance. In this work, we explore the impact of the several design parameters on the recognition accuracy by evaluating the system performance on the CHIME-2 and AURORA-2 databases.","keywords":"acoustic noise;speech recognition;adaptive noise dictionary design;feasible noise modeling;noise robust exemplar matching framework;multiple length exemplars linear combination;SR techniques;sparse representation formulation;noise-only training sequences;target noisy utterance;recognition;N-REM;noisy speech segments;Dictionaries;Speech;Training;Noise measurement;Hidden Markov models;Signal to noise ratio;Adaptation models;template matching;noise-robustness;automatic speech recognition;sparse representations;exemplar selection","doi":"10.1109/EUSIPCO.2015.7362670","issn":"2076-1465","month":"Aug","url":"https://www.eurasip.org/proceedings/eusipco/eusipco2015/papers/1570102603.pdf","bibtex":"@InProceedings{7362670,\n author = {E. Yilmaz and H. {Van hamme} and J. F. Gemmeke},\n booktitle = {2015 23rd European Signal Processing Conference (EUSIPCO)},\n title = {Adaptive noise dictionary design for noise robust exemplar matching of speech},\n year = {2015},\n pages = {1681-1685},\n abstract = {This paper investigates an adaptive noise dictionary design approach to achieve an effective and computationally feasible noise modeling for the noise robust exemplar matching (N-REM) framework. N-REM approximates noisy speech segments as a linear combination of multiple length exemplars in a sparse representation (SR) formulation. Compared to the previous SR techniques with a single overcomplete dictionary, N-REM uses smaller dictionaries containing considerably fewer noise exemplars. Hence, the noise exemplars have to be selected with care to accurately model the spectrotem-poral content of the actual noise conditions. For this purpose, in a previous work, we introduced a noise exemplar selection stage before performing recognition which extracts noise exemplars from a few noise-only training sequences chosen for each target noisy utterance. In this work, we explore the impact of the several design parameters on the recognition accuracy by evaluating the system performance on the CHIME-2 and AURORA-2 databases.},\n keywords = {acoustic noise;speech recognition;adaptive noise dictionary design;feasible noise modeling;noise robust exemplar matching framework;multiple length exemplars linear combination;SR techniques;sparse representation formulation;noise-only training sequences;target noisy utterance;recognition;N-REM;noisy speech segments;Dictionaries;Speech;Training;Noise measurement;Hidden Markov models;Signal to noise ratio;Adaptation models;template matching;noise-robustness;automatic speech recognition;sparse representations;exemplar selection},\n doi = {10.1109/EUSIPCO.2015.7362670},\n issn = {2076-1465},\n month = {Aug},\n url = {https://www.eurasip.org/proceedings/eusipco/eusipco2015/papers/1570102603.pdf},\n}\n\n","author_short":["Yilmaz, E.","Van hamme, H.","Gemmeke, J. F."],"key":"7362670","id":"7362670","bibbaseid":"yilmaz-vanhamme-gemmeke-adaptivenoisedictionarydesignfornoiserobustexemplarmatchingofspeech-2015","role":"author","urls":{"Paper":"https://www.eurasip.org/proceedings/eusipco/eusipco2015/papers/1570102603.pdf"},"keyword":["acoustic noise;speech recognition;adaptive noise dictionary design;feasible noise modeling;noise robust exemplar matching framework;multiple length exemplars linear combination;SR techniques;sparse representation formulation;noise-only training sequences;target noisy utterance;recognition;N-REM;noisy speech segments;Dictionaries;Speech;Training;Noise measurement;Hidden Markov models;Signal to noise ratio;Adaptation models;template matching;noise-robustness;automatic speech recognition;sparse representations;exemplar selection"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"inproceedings","biburl":"https://raw.githubusercontent.com/Roznn/EUSIPCO/main/eusipco2015url.bib","creationDate":"2021-02-13T17:31:52.467Z","downloads":0,"keywords":["acoustic noise;speech recognition;adaptive noise dictionary design;feasible noise modeling;noise robust exemplar matching framework;multiple length exemplars linear combination;sr techniques;sparse representation formulation;noise-only training sequences;target noisy utterance;recognition;n-rem;noisy speech segments;dictionaries;speech;training;noise measurement;hidden markov models;signal to noise ratio;adaptation models;template matching;noise-robustness;automatic speech recognition;sparse representations;exemplar selection"],"search_terms":["adaptive","noise","dictionary","design","noise","robust","exemplar","matching","speech","yilmaz","van hamme","gemmeke"],"title":"Adaptive noise dictionary design for noise robust exemplar matching of speech","year":2015,"dataSources":["eov4vbT6mnAiTpKji","knrZsDjSNHWtA9WNT"]}