Fusing Acoustic Feature Representations for Computational Paralinguistics Tasks

Kaya, Heysem; Karpov, Alexey

doi:10.21437/interspeech.2016-995

Cited by 26 publications

(31 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In line with experience on former challenge corpora [10,11,20], we investigate feature level fusion and two variants of score level fusion. The first is simple weighted fusion (SF) of scores, where the classifier confidence scores S A and S B are fused using weight γ ∈ [0, 1]:…”

Section: Fusionmentioning

confidence: 99%

“…In line with our recent experience on paralinguistic and multi-modal affective computing [10], we employ least squares based classifiers such as Kernel Extreme Learning Machines (KELM) and Partial Least Squares (PLS) regression based classifiers for modeling utterance-level feature representations. Furthermore, we employ their weighted versions to cope with the class imbalance problem, typically observed in challenge corpora [11].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

LSTM Based Cross-corpus and Cross-task Acoustic Emotion Recognition

Kaya

Fedotov

Yesilkanat³

et al. 2018

Interspeech 2018

Self Cite

View full text Add to dashboard Cite

Acoustic emotion recognition is a popular and central research direction in paralinguistic analysis, due its relation to a wide range of affective states/traits and manifold applications. Developing highly generalizable models still remains as a challenge for researchers and engineers, because of multitude of nuisance factors. To assert generalization, deployed models need to handle spontaneous speech recorded under different acoustic conditions compared to the training set. This requires that the models are tested for cross-corpus robustness. In this work, we first investigate the suitability of Long-Short-Term-Memory (LSTM) models trained with time-and space-continuously annotated affective primitives for cross-corpus acoustic emotion recognition. We next employ an effective approach to use the frame level valence and arousal predictions of LSTM models for utterance level affect classification and apply this approach on the ComParE 2018 challenge corpora. The proposed method alone gives motivating results both on development and test set of the Self-Assessed Affect Sub-Challenge. On the development set, the cross-corpus prediction based method gives a boost to performance when fused with top components of the baseline system. Results indicate the suitability of the proposed method for both time-continuous and utterance level cross-corpus acoustic emotion recognition tasks.

show abstract

Section: Fusionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

LSTM Based Cross-corpus and Cross-task Acoustic Emotion Recognition

Kaya

Fedotov

Yesilkanat³

et al. 2018

Interspeech 2018

Self Cite

View full text Add to dashboard Cite

show abstract

“…Условия данного конкурса представлены в статье организаторов [63], а результаты будут подведены в сентябре 2016 г. в ходе 17-й международной конференции INTERSPEECH-2016 в США. Турецко-российский коллектив авторов статьи также принимает участие в соревновании ComParE 2016 года [64]. …”

Section: заключениеunclassified

State-of-the-art tasks and achievements of paralinguistic speech analysis systems

Karpov¹,

Kaya²,

Salah³

2016

Naučno-teh. vestn. inf. tehnol. meh. opt.

View full text Add to dashboard Cite

“…Multi-modal decision level fusion is proposed in [8] for emotion recognition in the Wild [9]. Results presented in [10][11][12] show that the applied fusion methods improve the performance of the standalone detectors and provide systems capable of outperforming the baseline systems.…”

Section: Introductionmentioning

confidence: 99%

Implementing Fusion Techniques for the Classification of Paralinguistic Information

Vlasenko¹,

Sebastian²,

Kumar³

et al. 2018

Interspeech 2018

View full text Add to dashboard Cite

This work tests several classification techniques and acoustic features and further combines them using late fusion to classify paralinguistic information for the ComParE 2018 challenge. We use Multiple Linear Regression (MLR) with Ordinary Least Squares (OLS) analysis to select the most informative features for Self-Assessed Affect (SSA) sub-Challenge. We also propose to use raw-waveform convolutional neural networks (CNN) in the context of three paralinguistic sub-challenges. By using combined evaluation split for estimating codebook, we obtain better representation for Bag-of-Audio-Words approach. We preprocess the speech to vocalized segments to improve classification performance. For fusion of our leading classification techniques, we use weighted late fusion approach applied for confidence scores. We use two mismatched evaluation phases by exchanging the training and development sets, and this estimates the optimal fusion weight. Weighted late fusion provides better performance on development sets in comparison with baseline techniques. Raw-waveform techniques perform comparable to the baseline.

show abstract

Fusing Acoustic Feature Representations for Computational Paralinguistics Tasks

Cited by 26 publications

References 14 publications

LSTM Based Cross-corpus and Cross-task Acoustic Emotion Recognition

LSTM Based Cross-corpus and Cross-task Acoustic Emotion Recognition

State-of-the-art tasks and achievements of paralinguistic speech analysis systems

Implementing Fusion Techniques for the Classification of Paralinguistic Information

Contact Info

Product

Resources

About