2,360 Hits in 3.5 sec

Comparison of adaptation methods for GMM-SVM based speech emotion recognition

Jianbo Jiang, Zhiyong Wu, Mingxing Xu, Jia Jia, Lianhong Cai
2012 2012 IEEE Spoken Language Technology Workshop (SLT)  
To gain the accuracy rate of emotion distinction, adaptation algorithms that can be manipulated on short utterances are highly essential.  ...  Experiment results show that MLLR adaptation performs better for very short enrollment utterances (with the length shorter than 2s) while MAP adaptation is more effective for longer utterances.  ...  The universal background model (UBM) is trained from the database using Expectation-maximization (EM) algorithm.  ... 
doi:10.1109/slt.2012.6424234 dblp:conf/slt/JiangWXJC12 fatcat:deaywxdrivbxdfvrynsa3xravy

Speech-Driven Automatic Facial Expression Synthesis

Elif Bozkurt, Cigdem Eroglu Erdem, Engin Erzin, Tanju Erdem, Mehmet Ozkan, A. Murat Tekalp
2008 2008 3DTV Conference: The True Vision - Capture, Transmission and Display of 3D Video  
GMM classifier has the best overall recognition rate 82.85% when cepstral features with delta and acceleration coefficients are used.  ...  Then, we classify the seven emotions in the dataset with two different classifiers: Gaussian mixture models (GMMs) and Hidden Markov Models (HMMs).  ...  However, in comparison to GMM results, they are lower (37.59% in average). For fair comparison to GMMs, we use grammar model that matches each utterance only to one emotion.  ... 
doi:10.1109/3dtv.2008.4547861 fatcat:rc5gqmljtbh2dl4bjowhwl2kgi

Cross-layer energy optimization under image quality constraints for wireless image transmissions

Na Yang, Ilker Demirkol, Wendi Heinzelman
2012 2012 8th International Wireless Communications and Mobile Computing Conference (IWCMC)  
for the LDC database Figure 2 . 5 : 25 GPE rate of the different algorithms for the LDC database [1] , averaged over all eight types of noise.  ...  Figure 2 . 7 : 27 GPE rate of the different algorithms for the KEELE database [4] , averaged over all eight types of noise.  ... 
doi:10.1109/iwcmc.2012.6314342 dblp:conf/iwcmc/YangDH12 fatcat:fpjsdzsu25dkbghfas5uonty3a

Impact of Feature Selection Algorithm on Speech Emotion Recognition Using Deep Convolutional Neural Network

Misbah Farooq, Fawad Hussain, Naveed Khan Baloch, Fawad Riasat Raja, Heejung Yu, Yousaf Bin Zikria
2020 Sensors  
In this study, the benefits of a deep convolutional neural network (DCNN) for SER are explored.  ...  For the classification of emotions, we utilize support vector machines, random forests, the k-nearest neighbors algorithm, and neural network classifiers.  ...  A spotting algorithm that examines spoken utterances for emotional keywords or phrases was proposed in [9] . The keyword spotting algorithm deals with the detection of keywords in the utterances.  ... 
doi:10.3390/s20216008 pmid:33113907 fatcat:wvnjjivr5jdfnptc2fdrjtxbou

Automatic refinement of an expressive speech corpus assembling subjective perception and automatic classification

Ignasi Iriondo, Santiago Planet, Joan-Claudi Socoró, Elisa Martínez, Francesc Alías, Carlos Monzo
2009 Speech Communication  
An expressive speech corpus in Spanish, designed and recorded for speech synthesis purposes, has been used to test the presented proposal.  ...  The content which most closely matches the subjective classification remains in the resulting corpus.  ...  The SES database included four emotions (sadness, happiness, anger and surprise) as well as a neutral expression.  ... 
doi:10.1016/j.specom.2008.12.001 fatcat:zxccupc4ffgbjc52o34dunv4ky

Facial Expression Recognition by Inter-class Relational Learning

Yizhen Chen, Haifeng Hu
2019 IEEE Access  
Then, two extracted features are integrated with a random ratio so that a hybrid feature can be attained. An attention module is proposed to assign a weight for each pixel of the hybrid feature.  ...  The IcRL method has been evaluated on five public expression databases: CK+, JAFFE, TFEID, BAUM-2i, and FER2013.  ...  TABLE 2 . 2 Average accuracy on the CK+ database (%). TABLE 3 . 3 Average accuracy on the JAFFE database (%). TABLE 5 . 5 Average accuracy on BAUM-2i and FER2013 (%).  ... 
doi:10.1109/access.2019.2928983 fatcat:xlgxm77hcvdatjjgdgty46zh5y

Class-level spectral features for emotion recognition

Dmitri Bitouk, Ragini Verma, Ani Nenkova
2010 Speech Communication  
, unstressed vowels and consonants in the utterance.  ...  , unstressed vowels and consonants in the utterance.  ...  Jiahong Yuan for providing us with the code and English acoustic models for forced alignment.  ... 
doi:10.1016/j.specom.2010.02.010 pmid:23794771 pmcid:PMC3686526 fatcat:oifrkjr325hgfcodsoyripwghq

Capturing global spatial patterns for distinguishing posed and spontaneous expressions

Shangfei Wang, Chongliang Wu, Qiang Ji
2016 Computer Vision and Image Understanding  
Furthermore, we propose efficient inference algorithm by extending annealing importance sampling to RBM with continuous visible units for calculating partition function of RBMs.  ...  Experimental results on benchmark databases demonstrate the effectiveness of the proposed approach in modelling global spatial patterns as well as its superior posed and spontaneous expression distinction  ...  The procedure of AIS algorithm displayed in Algorithm 2.  ... 
doi:10.1016/j.cviu.2015.08.007 fatcat:a75khkkt7zdafleaujw3tsgnhm

A Fully Dynamic Algorithm for k-Regret Minimizing Sets [article]

Yanhao Wang, Yuchen Li, Raymond Chi-Wing Wong, Kian-Lee Tan
2020 arXiv   pre-print
To address this issue, we propose the first fully-dynamic algorithm for the k-RMS problem that can efficiently provide the up-to-date result w.r.t. any insertion and deletion in the database with a provable  ...  Although the k-RMS problem has been extensively studied in the literature, existing methods are designed for the static setting and cannot maintain the result efficiently when the database is updated.  ...  ACKNOWLEDGMENT We thank anonymous reviewers for their helpful comments to improve this research. Yanhao Wang has been supported by the MLDB project of Academy of Finland (decision number: 322046).  ... 
arXiv:2005.14493v3 fatcat:pqhzgqzrorc5jksqpcntub5tw4

A Generic Design of Driver Drowsiness and Stress Recognition Using MOGA Optimized Deep MKL-SVM

Kwok Tai Chui, Miltiadis D. Lytras, Ryan Wen Liu
2020 Sensors  
Researchers are in full swing to develop various algorithms for both drowsiness and stress recognition.  ...  Results reveal that the proposed algorithm achieves an average sensitivity of 99%, specificity of 98.3% and area under the receiver operating characteristic curve (AUC) of 97.1% for driver drowsiness recognition  ...  O2 maximizes the specificity, which is the ratio of true negative (TN) and number of negative samples (Nn).  ... 
doi:10.3390/s20051474 pmid:32156100 pmcid:PMC7085776 fatcat:7x3pq7tu4bdx3mdgqzfdvjcuam

Expressive Music Performance Modeling

Andreas Neocleous, Rafael Ramírez
2010 Zenodo  
In particular, we investigated how professional musicians encode emotions, such as happiness, sadness, anger, fear and sweetness, in violin and saxophone audio performances.  ...  Machine learning approaches to modelling emotions in music performances were investigated and presented in this thesis.  ...  and bow acceleration).  ... 
doi:10.5281/zenodo.3753047 fatcat:nu6p5hbfsnggnirxwwjrxpqhzq

Predicting sex as a soft-biometrics from device interaction swipe gestures

Oscar Miguel-Hurtado, Sarah V. Stevenage, Chris Bevan, Richard Guest
2016 Pattern Recognition Letters  
Following research studies undertaken in similar modalities such as keystroke and mouse usage biometrics, the present work proposes the use of swipe gesture data for the prediction of soft-biometrics,  ...  Within this analysis, the BestFirst feature selection technique and classification algorithms (naïve Bayes, logistic regression, support vector machine and decision tree) have been tested.  ...  The authors would like to thank colleagues on this grant for helpful comments on this work.  ... 
doi:10.1016/j.patrec.2016.04.024 fatcat:hw23mtgv2ffctci5ronulgs24a

Ensemble Learning with Attention-Integrated Convolutional Recurrent Neural Network for Imbalanced Speech Emotion Recognition

Xusheng Ai, Victor S. Sheng, Wei Fang, Charles X. Ling, Chunhua Li
2020 IEEE Access  
That is, we have 10 samples in all for each database. We save the optimal model which achieves the highest UAR.  ...  Researchers have proposed methods or algorithms to address the imbalance issue for standard machine learning algorithms, and some of them are also valid for deep learning algorithms [11] , [12] .  ... 
doi:10.1109/access.2020.3035910 fatcat:rgf2mnn3vnhifhez6e72a5fihi

Smith-Waterman Acceleration in Multi-GPUs: A Performance per Watt Analysis [chapter]

Jesús Pérez Serrano, Edans Flavius De Oliveira Sandes, Alba Cristina Magalhaes Alves de Melo, Manuel Ujaldón
2017 Lecture Notes in Computer Science  
Speed-up factors and energy consumption are monitored on different stages of the algorithm with the goal of identifying advantageous scenarios to maximize acceleration and minimize power consumption.  ...  We present a performance per watt analysis of CUDAlign 4.0, a parallel strategy to obtain the optimal alignment of huge DNA sequences in multi-GPU platforms using the exact Smith-Waterman method.  ...  As in the original SW algorithm, time and space complexities of the Gotoh algorithm are quadratic.  ... 
doi:10.1007/978-3-319-56154-7_46 fatcat:q4mfjtje4nehpc67yvmn7w47jy

Gender Estimation Based on Smile-Dynamics

Antitza Dantcheva, Francois Bremond
2017 IEEE Transactions on Information Forensics and Security  
Most commonly, the underlying algorithms analyze the facial appearance for clues of gender.  ...  In addition, we fuse proposed dynamics-based approach with state-ofthe-art appearance based algorithms, predominantly improving appearance-based gender estimation performance.  ...  (D + ) (D + )+ (D − ) (D − ) (D + )+ (D − ) Maximal Speed max(V + ) max(V − ) Mean Speed mean(V + ) mean(V − ) Maximum Acceleration max(A + ) max(A − ) Mean Acceleration mean(A + ) mean(A − ) Net Ampl  ... 
doi:10.1109/tifs.2016.2632070 fatcat:goqe3e727rhanmsyp3elyyjrdu
« Previous Showing results 1 — 15 out of 2,360 results