Fisher vectors with cascaded normalization for paralinguistic analysis

Kaya, Heysem; Karpov, Alexey; Salah, Albert Ali

doi:10.21437/interspeech.2015-193

Cited by 39 publications

(11 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The UAR of the baseline training partition, set using LOSO-CV, was 61.3%, and the test set UAR 65.9%. The winning entry to the challenge was obtained using a system based on Fisher vector encoding of extracted features and cascaded normalisation to account for speaker and phonetic variability (Kaya et al, 2015). This approach achieved a test-set UAR of 83.1%.…”

Section: Ihearu-eat Databasementioning

confidence: 99%

“…The UARs for the training and test partition for this system were 35.2% and 32.8%, respectively. Despite the additional information stream provided by the video data, no entrants to the 2018 challenge were able to better the UAR of 83.1% obtained in Kaya et al (2015). A detailed summary of the EAT challenge can be found in Schuller and Schuller (2020).…”

Section: Ihearu-eat Databasementioning

confidence: 99%

See 1 more Smart Citation

Representation transfer learning from deep end-to-end speech recognition networks for the classification of health states from speech

Sertolli

Ren

Schuller

et al. 2021

Computer Speech & Language

View full text Add to dashboard Cite

Section: Ihearu-eat Databasementioning

confidence: 99%

Section: Ihearu-eat Databasementioning

confidence: 99%

Representation transfer learning from deep end-to-end speech recognition networks for the classification of health states from speech

Sertolli

Ren

Schuller

et al. 2021

Computer Speech & Language

View full text Add to dashboard Cite

“…Despite the fact that just a handful of studies use FV in speech processing (e.g. for categorizing audio-signals as speech, music and others [29], for speaker verification [30,31], and for determining the food type from eating sounds [32]), we think that FV can be harnessed to improve classification performance in audio processing.…”

Section: Fisher Vectorsmentioning

confidence: 99%

They are wearing a mask! Identification of Subjects Wearing a Surgical Mask from their Speech by means of x-vectors and Fisher Vectors

Egas-López

2020

Preprint

View full text Add to dashboard Cite

Challenges based on Computational Paralinguistics in the IN-TERSPEECH Conference have always had a good reception among the attendees, owing to its competitive academic and research demands. This year, the INTERSPEECH 2020 Computational Paralinguistics Challenge offers three different problems; here, the Mask Sub-Challenge is of specific interest. This challenge involves the classification of speech recorded from subjects while wearing a surgical mask. In this study, to address the above mentioned problem we employ two different types of feature extraction methods. The x-vectors embeddings, which is the current state-of-the-art approach for Speaker Recognition; and the Fisher Vector (FV), that is a method originally intended for Image Recognition, but here we utilize it to discriminate utterances. These approaches employ distinct frame-level representations: MFCC and PLP. Using Support Vector Machines (SVM) as the classifier, we perform a technical comparison between the performances of the FV encodings and the x-vector embeddings for this particular classification task. We find that the Fisher vector encodings provide better representations of the utterances than the x-vectors do for this specific dataset. Moreover, we show that a fusion of our best configurations outperforms all the baseline scores of the Mask Sub-Challenge.

show abstract

“…However, the constituents and the combination rules of the ensemble systems must be selected with care. Kaya and colleagues previously applied the Fisher Vector (FV) encoding of acoustic Low-Level Descriptors (LLD) to several paralinguistic tasks including recognition of native language and sincerity [7], as well as classification of snoring types [8] and eating conditions [9]. In this work, we use a similar FV encoding for the representation of acoustic features.…”

Section: Introductionmentioning

confidence: 99%

Is Everything Fine, Grandma? Acoustic and Linguistic Modeling for Robust Elderly Speech Emotion Recognition

Sogancioglu¹,

Verkholyak²,

Kaya³

et al. 2020

Interspeech 2020

Self Cite

View full text Add to dashboard Cite

Acoustic and linguistic analysis for elderly emotion recognition is an under-studied and challenging research direction, but essential for the creation of digital assistants for the elderly, as well as unobtrusive telemonitoring of elderly in their residences for mental healthcare purposes. This paper presents our contribution to the INTERSPEECH 2020 Computational Paralinguistics Challenge (ComParE) -Elderly Emotion Sub-Challenge, which is comprised of two ternary classification tasks for arousal and valence recognition. We propose a bimodal framework, where these tasks are modeled using stateof-the-art acoustic and linguistic features, respectively. In this study, we demonstrate that exploiting task-specific dictionaries and resources can boost the performance of linguistic models, when the amount of labeled data is small. Observing a high mismatch between development and test set performances of various models, we also propose alternative training and decision fusion strategies to better estimate and improve the generalization performance.

show abstract

Fisher vectors with cascaded normalization for paralinguistic analysis

Cited by 39 publications

References 30 publications

Representation transfer learning from deep end-to-end speech recognition networks for the classification of health states from speech

Representation transfer learning from deep end-to-end speech recognition networks for the classification of health states from speech

They are wearing a mask! Identification of Subjects Wearing a Surgical Mask from their Speech by means of x-vectors and Fisher Vectors

Is Everything Fine, Grandma? Acoustic and Linguistic Modeling for Robust Elderly Speech Emotion Recognition

Contact Info

Product

Resources

About