Faculty Publications

Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736

Publications by NITK Faculty

Browse

Author: search.filters.author.Kalluri, S.B.Has files: No

Search Results

Now showing 1 - 7 of 7
  • No Thumbnail Available
    Item
    Estimating multiple physical parameters from speech data
    (IEEE Computer Society help@computer.org, 2016) Kalluri, S.B.; Vijayakumar, A.; Vijayasenan, D.; Singh, R.
    In this work, we explore prediction of different physical parameters from speech data. We aim to predict shoulder size and waist size of people from speech data in addition to the conventional height and weight parameters. A data-set with this information is created from 207 volunteers. A bag of words representation based on log magnitude spectrum is used as features. A support vector regression predicts the physical parameters from the bag of the words representation. The system is able to achieve a root mean square error of 6.6 cm for height estimation, 2.6cm for shoulder size, 7.1cm for waist size and 8.9 kg for weight estimation. The results of height estimation is on par with state of the art results. © 2016 IEEE.
  • No Thumbnail Available
    Item
    Study of Wireless Channel Effects on Audio Forensics
    (Institute of Electrical and Electronics Engineers Inc., 2018) Vijayasenan, D.; Kalluri, S.B.; Sreekanth, K.; Issac, A.
    In this work, we try to study the effect of a wireless channel on physical parameter prediction based on speech data. Speech data from 207 speakers along with corresponding speaker's height and weight is collected. A three path Rayleigh fading channel with typical values of Doppler shift, path gain and path delay is utilized to create the mobile channel output audio. A Bag of Words (BoW) representation based on log magnitude spectrum is used as features. Support Vector Regression (SVR) predicts the physical parameter of the speaker from the BoW representation. The proposed system is able to achieve a Root Mean Square Error (RMSE) of 6.6 cm for height estimation and 8.9 Kg for weight estimation for clean speech. The effect of Rayleigh channel increase the RMSE values to 8.17 cm and 11.84 Kg respectively for height and weight. © 2016 IEEE.
  • No Thumbnail Available
    Item
    A Deep Neural Network Based End to End Model for Joint Height and Age Estimation from Short Duration Speech
    (Institute of Electrical and Electronics Engineers Inc., 2019) Kalluri, S.B.; Vijayasenan, D.; Ganapathy, S.
    Automatic height and age prediction of a speaker has a wide variety of applications in speaker profiling, forensics etc. Often in such applications only a few seconds of speech data is available to reliably estimate the speaker parameters. Traditionally, age and height were predicted separately using different estimation algorithms. In this work, we propose a unified DNN architecture to predict both height and age of a speaker for short durations of speech. A novel initialization scheme for the deep neural architecture is introduced, that avoids the requirement for a large training dataset. We evaluate the system on TIMIT dataset where the mean duration of speech segments is around 2.5s. The DNN system is able to improve the age RMSE by at least 0.6 years as compared to a conventional support vector regression system trained on Gaussian Mixture Model mean supervectors. The system achieves an RMSE error of 6.85 and 6.29 cm for male and female height prediction. In case of age estimation, the RMSE errors are 7.60 and 8.63 years for male and female respectively. Analysis of shorter speech segments reveals that even with 1 second speech input the performance degradation is at most 3% compared to the full duration speech files. © 2019 IEEE.
  • No Thumbnail Available
    Item
    Nisp: A multi-lingual multi-accent dataset for speaker profiling
    (Institute of Electrical and Electronics Engineers Inc., 2021) Kalluri, S.B.; Vijayasenan, D.; Ganapathy, S.; Rajan, M.; Krishnan, P.
    Many commercial and forensic applications of speech demand the extraction of information about the speaker characteristics, which falls into the broad category of speaker profiling. The speaker characteristics needed for profiling include physical traits of the speaker like height, age, and gender of the speaker along with the native language of the speaker. Many of the datasets available have only partial information for speaker profiling. In this paper, we attempt to overcome this limitation by developing a new dataset which has speech data from five different Indian languages along with English. The metadata information for speaker profiling applications like linguistic information, regional information, and physical characteristics of a speaker are also collected. We call this dataset as NITK-IISc Multilingual Multi-accent Speaker Profiling (NISP) dataset. The description of the dataset, potential applications, and baseline results for speaker profiling on this dataset are provided in this paper. © 2021 IEEE.
  • No Thumbnail Available
    Item
    COVID-19 detection from spectral features on the DiCOVA dataset
    (International Speech Communication Association, 2021) Ritwik, K.V.S.; Kalluri, S.B.; Vijayasenan, D.
    In this paper we investigate the cues of COVID-19 on sustained phonation of Vowel-/i/, deep breathing and number counting data of the DiCOVA dataset. We use an ensemble of classifiers trained on different features, namely, super-vectors, formants, harmonics and MFCC features. We fit a two-class Weighted SVM classifier to separate the COVID-19 audio from Non-COVID-19 audio. Weighted penalties help mitigate the challenge of class imbalance in the dataset. The results are reported on the stationary (breathing, Vowel-/i/) and nonstationary( counting data) data using individual and combination of features on each type of utterance. We find that the Formant information plays a crucial role in classification. The proposed system resulted in an AUC score of 0.734 for cross validation, and 0.717 for evaluation dataset. © © 2021 ISCA.
  • No Thumbnail Available
    Item
    The Second DISPLACE Challenge: DIarization of SPeaker and LAnguage in Conversational Environments
    (International Speech Communication Association, 2024) Kalluri, S.B.; Singh, P.; Roy Chowdhuri, P.; Kulkarni, A.; Baghel, S.; Hegde, P.; Sontakke, S.; Deepak, K.T.; Mahadeva Prasanna, S.R.; Vijayasenan, D.; Ganapathy, S.
    The DIarization of SPeaker and LAnguage in Conversational Environments (DISPLACE) 2024 challenge is the second in the series of DISPLACE challenges, which involves tasks of speaker diarization (SD) and language diarization (LD) on a challenging multilingual conversational speech dataset. In the DISPLACE 2024 challenge, we also introduced the task of automatic speech recognition (ASR) on this dataset. The dataset containing 158 hours of speech, consisting of both supervised and unsupervised mono-channel far-field recordings, was released for LD and SD tracks. Further, 12 hours of close-field mono-channel recordings were provided for the ASR track conducted on 5 Indian languages. The details of the dataset, baseline systems and the leader board results are highlighted in this paper. We have also compared our baseline models and the team's performances on evaluation data of DISPLACE-2023 to emphasize the advancements made in this second version of the challenge. © 2024 International Speech Communication Association. All rights reserved.
  • No Thumbnail Available
    Item
    Automatic speaker profiling from short duration speech data
    (Elsevier B.V., 2020) Kalluri, S.B.; Vijayasenan, D.; Ganapathy, S.
    Many paralinguistic applications of speech demand the extraction of information about the speaker characteristics from as little speech data as possible. In this work, we explore the estimation of multiple physical parameters of the speaker from the short duration of speech in a multilingual setting. We explore different feature streams for age and body build estimation derived from the speech spectrum at different resolutions, namely – short-term log-mel spectrogram, formant features and harmonic features of the speech. The statistics of these features over the speech recording are used to learn a support vector regression model for speaker age and body build estimation. The experiments performed on the TIMIT dataset show that each of the individual features is able to achieve results that outperform previously published results in height and age estimation. Furthermore, the estimation errors from these different feature streams are complementary, which allows the combination of estimates from these feature streams to further improve the results. The combined system from short audio snippets achieves a performance of 5.2 cm, and 4.8 cm in Mean Absolute Error (MAE) for male and female respectively for height estimation. Similarly in age estimation the MAE is of 5.2 years, and 5.6 years for male, and female speakers respectively. We also extend the same physical parameter estimation to other body build parameters like shoulder width, waist size and weight along with height on a dataset we collected for speaker profiling. The duration analysis of the proposed scheme shows that the state of the art results can be achieved using only around 1–2 s of speech data. To the best of our knowledge, this is the first attempt to use a common set of features for estimating the different physical traits of a speaker. © 2020 Elsevier B.V.