Acoustic-Phonetic Feature Based Dialect Identification in Hindi Speech

Sinha, Shweta; Jain, Aruna; Agrawal, S.

doi:10.21307/ijssis-2017-757

Cited by 12 publications

(11 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Mô hình hỗn hợp Gauss đa thể hiện (Gaussian Mixture Model: GMM) đã được sử dụng trong các nghiên cứu về nhận dạng người nói [3], định danh phương ngữ tiếng Anh [4], tiếng Trung [5], tiếng Thái [6], tiếng Hindi [7], tiếng Việt [8], nhận dạng ngôn ngữ [9], [10]. Supervectors cũng được sử dụng trong nghiên cứu nhận dạng phương ngữ và cho kết quả khả quan [11].…”

Section: B Nhận Dạng Phương Ngữ Tiếng Việt Dùng Mô Hình Gmm Với Mfccunclassified

Cải Thiện Hiệu Năng Hệ Thống Nhận Dạng Tiếng Việt Với Thông Tin Về Phương Ngữ

Hùng¹,

Loan²,

Quang³

et al. 2017

Fair - Nghiên Cứu Cơ Bản Và Ứng Dụng Công Nghệ Thông Tin - 2016

View full text Add to dashboard Cite

Section: B Nhận Dạng Phương Ngữ Tiếng Việt Dùng Mô Hình Gmm Với Mfccunclassified

Cải Thiện Hiệu Năng Hệ Thống Nhận Dạng Tiếng Việt Với Thông Tin Về Phương Ngữ

Hùng¹,

Loan²,

Quang³

et al. 2017

Fair - Nghiên Cứu Cơ Bản Và Ứng Dụng Công Nghệ Thông Tin - 2016

View full text Add to dashboard Cite

“…The first two formants are the most important because they decide the speech quality [14]. Formants and their bandwidths have been used for a lot of research on speech processing such as accent identification [15][16][17], speech recognition [18], speaker identification [19], study on genders and ethnical accents [20][21][22], dialect identification [4,[23][24][25].…”

Section: Selection Of the Number Of Coefficients Mfccmentioning

confidence: 99%

Automatic Identification of Vietnamese Dialects

Hùng

Loan

Quang

2016

JCC

View full text Add to dashboard Cite

The dialect identification was studied for many languages over the world nevertheless the research on signal processing for Vietnamese dialects is still limited and there were not many published works. There are many different dialects for Vietnamese. The influence of dialectal features on speech recognition systems is important. If the information about dialects is known during speech recognition process, the performance of recognition systems will be better because the corpus of these systems is normally organized according to different dialects. This paper will present the combination of MFCC coefficients and fundamental frequency features of Vietnamese for dialectal identification based on GMM. The experiment result for the dialect corpus of Vietnamese shows that the performance of dialectal identification is increased from 59% for the case using only MFCC coefficients to 71% for the case using MFCC coefficients and the information of fundamental frequency.

show abstract

“…Examples are Mel-frequency cepstral coefficients (MFCCs; e.g. [3,4,5,10]), signal energy (e.g. [4,10]), Perceptual Linear Prediction coefficients (e.g.…”

Section: Introductionmentioning

confidence: 99%

“…[3,4,5,10]), signal energy (e.g. [4,10]), Perceptual Linear Prediction coefficients (e.g. [4,10,11]), voicing probability (e.g.…”

Section: Introductionmentioning

confidence: 99%

Styrian Dialect Classification: Comparing and Fusing Classifiers Based on a Feature Selection Using a Genetic Algorithm

2019

View full text Add to dashboard Cite

Many classifiers struggle when confronted with a high dimensional feature space like in the data sets provided for the Interspeech ComParE challenge. This is because most features do not significantly contribute to the prediction. To alleviate this problem, we propose a feature selection based on a Genetic Algorithm (GA) that uses an SVM as the fitness function. We show that this yields a reduced subset (1) which results in an Unweighted Average Recall (UAR) that beats the challenge baseline on the development set for the 3-class classification problem. Further, we extract an additional per-phoneme feature set, where the features are inspired by the ComParE features. On this set the same GA-based feature selection is performed and the resulting set is used for training in isolation (2) and in combination with the aforementioned reduced challenge features (3). Five classifiers were tested on the three subsets, namely SVMs, DNNs, GBMs, RFs, and regularized regression. All classifiers achieved a UAR above the baseline on all three sets. The best performance on set (1) was achieved by an SVM using an RBF kernel and on sets (2) and (3) by a fusion of classifiers.

show abstract

Acoustic-Phonetic Feature Based Dialect Identification in Hindi Speech

Cited by 12 publications

References 26 publications

Cải Thiện Hiệu Năng Hệ Thống Nhận Dạng Tiếng Việt Với Thông Tin Về Phương Ngữ

Cải Thiện Hiệu Năng Hệ Thống Nhận Dạng Tiếng Việt Với Thông Tin Về Phương Ngữ

Automatic Identification of Vietnamese Dialects

Styrian Dialect Classification: Comparing and Fusing Classifiers Based on a Feature Selection Using a Genetic Algorithm

Contact Info

Product

Resources

About