Faculty Publications
Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736
Publications by NITK Faculty
Browse
4 results
Search Results
Item Automatic text-independent Kannada dialect identification system(Springer Verlag service@springer.de, 2019) Chittaragi, N.B.; Limaye, A.; Chandana, N.T.; Annappa, B.; Koolagudi, S.G.This paper proposes a dialect identification system for the Kannada language. A system that can automatically identify the dialects of the language being spoken has a wide variety of applications. However, not many Automatic Speech Recognition (ASR) and dialect identification tasks are carried out in majority of the Indian languages. Further, there are only a few good quality annotated audio datasets available. In this paper, a new dataset for 5 spoken dialects of the Kannada language is introduced. Spectral and prosodic features have captured the most prominent features for recognition of Kannada dialects. Support Vector Machine (SVM) and neural networks algorithms are used for modeling text-independent recognition system. A neural network model that attempts for identification dialects based on sentence level cues has also been built. Hyper-parameters for SVM and neural network models are chosen using grid search. Neural network models have outperformed SVMs when complete utterances are considered. © Springer Nature Singapore Pte Ltd. 2019.Item Kannada Dialect Identification from Case-Based Word Utterances Using Gradient Boosting Algorithm(Springer Science and Business Media Deutschland GmbH, 2022) Chittaragi, N.B.; Koolagudi, S.G.Dialects or accents constitute the grammatical variations along with phonological and lexical changes those are commonly observed in the usage of a language with minor and subtle differences. Dialectal variations existing among dialects are mainly due to unique speaking patterns followed among the group of speakers. The dialect processing systems are essential in the development of automatic speech recognition systems (ASRs) for regional and resource-constrained languages in the country like India. Since India is with rich diversity in languages. In this paper, a language-dependent dialect identification system is proposed for Kannada language from words especially with the Kannada language-specific case (Vibhakthi Prathyayas) information. Special morphological operations that exist in the Kannada language in terms of various cases commonly called as a grammatical function of a noun or pronoun. These word utterances are used for the classification of five dialects of Kannada. This is a novel idea to use the smaller word utterances that consist of dialect-specific information representing the unique characteristics. In this paper, case-based word utterance dataset is prepared by considering five Kannada dialects from Kannada Dialect Speech Corpus (KDSC). Dynamic and static prosodic features are extracted to capture dialectal variations. Addition to these features, spectral MFCC features are also considered for evaluation of differences among dialects from these word-level units. Initially, multi-class Support vector machine (SVM) technique is used and later effective extreme gradient boosting (XGB) ensemble algorithms are used for the development of an automatic Kannada dialect recognition system. The research findings have demonstrated the words with case information convey dialect specific linguistic cues effectively. The combination of dynamic and static prosodic cues has a significant effect on the characterization of dialects along with spectral features. © 2022, Springer Nature Switzerland AG.Item Dialect Identification using Chroma-Spectral Shape Features with Ensemble Technique(Academic Press, 2021) Chittaragi, N.B.; Koolagudi, S.G.The present work proposes a text-independent dialect identification system. Generally, dialects of a language exhibit varying pronunciation styles followed in a specific geographical region. In this paper, chroma features familiar with music-related systems are employed for identification of dialects. In addition, eight significant spectral shape related features from short term spectra are computed and combined along with chroma features and named as chroma-spectral shape features. Chroma features try to aggregate spectral information and attempt to encapsulate the evidential variations, concerning timbre, correlated melody, rhythmic, and intonation patterns found prominently among dialects of few languages. The effectiveness of the proposed features and approach is evaluated on five prominent Kannada dialects spoken in Karnataka, India and internationally known standard Intonation Variation in English (IViE) dataset with nine British English dialects. Discriminative models such as, single classifier based Support Vector Machine (SVM) and ensemble based support vector machines (ESVM) are employed for classification. The proposed features have shown better performance over state-of-the-art i-vector features on both datasets. The highest recognition performance of 95.6% and 97.52% are achieved in the cases of Kannada and IViE dialect datasets respectively using ESVM. Proposed features have also demonstrated robust performance with small sized (limited data) audio clips even in noisy conditions. © 2021 Elsevier LtdItem Automatic diagnosis of COVID-19 related respiratory diseases from speech(Springer, 2023) Shekhar, K.; Chittaragi, N.B.; Koolagudi, S.G.In this work, an attempt is made to propose an intelligent and automatic system to recognize COVID-19 related illnesses from mere speech samples by using automatic speech processing techniques. We used a standard crowd-sourced dataset which was collected by the University of Cambridge through a web based application and an android/iPhone app. We worked on cough and breath datasets individually, and also with a combination of both the datasets. We trained the datasets on two sets of features, one consisting of only standard audio features such as spectral and prosodic features and one combining excitation source features with standard audio features extracted, and trained our model on shallow classifiers such as ensemble classifiers and SVM classification methods. Our model has shown better performance on both breath and cough datasets, but the best results in each of the cases was obtained through different combinations of features and classifiers. We got our best result when we used only standard audio features, and combined both cough and breath data. In this case, we achieved an accuracy of 84% and an Area Under Curve (AUC) score of 84%. Intelligent systems have already started to make a mark in medical diagnosis, and this type of study can help better the health system by providing much needed assistance to the health workers. © 2023, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.