Adaptive bands filter bank optimized by genetic algorithm for robust speech recognition system

Perceptual auditory filter banks such as Bark-scale filter bank are widely used as front-end processing in speech recognition systems.However,the problem of the design of optimized filter banks that provide higher accuracy in recognition tasks is still open.Owing to spectral analysis in feature extraction,an adaptive bands filter bank (ABFB) is presented.The design adopts flexible bandwidths and center frequencies for the frequency responses of the filters and utilizes genetic algorithm (GA) to optimize the design parameters.The optimization process is realized by combining the front-end filter bank with the back-end recognition network in the performance evaluation loop.The deployment of ABFB together with zero-crossing peak amplitude (ZCPA) feature as a front process for radial basis function (RBF) system shows significant improvement in robustness compared with the Bark-scale filter bank.In ABFB,several sub-bands are still more concentrated toward lower frequency but their exact locations are determined by the performance rather than the perceptual criteria.For the ease of optimization,only symmetrical bands are considered here,which still provide satisfactory results.

Robust Speech Recognition System Using Conventional and Hybrid Features of MFCC,LPCC,PLP,RASTA-PLP and Hidden Markov Model Classifier in Noisy Conditions

In recent years, the accuracy of speech recognition (SR) has been one of the most active areas of research. Despite that SR systems are working reasonably well in quiet conditions, they still suffer severe performance degradation in noisy conditions or distorted channels. It is necessary to search for more robust feature extraction methods to gain better performance in adverse conditions. This paper investigates the performance of conventional and new hybrid speech feature extraction algorithms of Mel Frequency Cepstrum Coefficient (MFCC), Linear Prediction Coding Coefficient (LPCC), perceptual linear production (PLP), and RASTA-PLP in noisy conditions through using multivariate Hidden Markov Model (HMM) classifier. The behavior of the proposal system is evaluated using TIDIGIT human voice dataset corpora, recorded from 208 different adult speakers in both training and testing process. The theoretical basis for speech processing and classifier procedures were presented, and the recognition results were obtained based on word recognition rate.

Study on Unequal Error Protection for Distributed Speech Recognition System

The unequal error protection （UEP） is applied in distributed speech recognition （DSR） system and three schemes are proposed. All of these three schemes are evaluated on the GSM simulating platform for recognizing mandarin digit strings and compared with the equal error protection （EEP） scheme. Experiments show that UEP can protect the data transmitted in DSR system more effectively, which results in a higher word accurate rate of DSR system.

Audio-Text Multimodal Speech Recognition via Dual-Tower Architecture for Mandarin Air Traffic Control Communications

In air traffic control communications (ATCC), misunderstandings between pilots and controllers could result in fatal aviation accidents. Fortunately, advanced automatic speech recognition technology has emerged as a promising means of preventing miscommunications and enhancing aviation safety. However, most existing speech recognition methods merely incorporate external language models on the decoder side, leading to insufficient semantic alignment between speech and text modalities during the encoding phase. Furthermore, it is challenging to model acoustic context dependencies over long distances due to the longer speech sequences than text, especially for the extended ATCC data. To address these issues, we propose a speech-text multimodal dual-tower architecture for speech recognition. It employs cross-modal interactions to achieve close semantic alignment during the encoding stage and strengthen its capabilities in modeling auditory long-distance context dependencies. In addition, a two-stage training strategy is elaborately devised to derive semantics-aware acoustic representations effectively. The first stage focuses on pre-training the speech-text multimodal encoding module to enhance inter-modal semantic alignment and aural long-distance context dependencies. The second stage fine-tunes the entire network to bridge the input modality variation gap between the training and inference phases and boost generalization performance. Extensive experiments demonstrate the effectiveness of the proposed speech-text multimodal speech recognition method on the ATCC and AISHELL-1 datasets. It reduces the character error rate to 6.54% and 8.73%, respectively, and exhibits substantial performance gains of 28.76% and 23.82% compared with the best baseline model. The case studies indicate that the obtained semantics-aware acoustic representations aid in accurately recognizing terms with similar pronunciations but distinctive semantics. The research provides a novel modeling paradigm for semantics-aware speech recognition in air traffic control communications, which could contribute to the advancement of intelligent and efficient aviation safety management.

Multi-Objective Equilibrium Optimizer for Feature Selection in High-Dimensional English Speech Emotion Recognition

Speech emotion recognition(SER)uses acoustic analysis to find features for emotion recognition and examines variations in voice that are caused by emotions.The number of features acquired with acoustic analysis is extremely high,so we introduce a hybrid filter-wrapper feature selection algorithm based on an improved equilibrium optimizer for constructing an emotion recognition system.The proposed algorithm implements multi-objective emotion recognition with the minimum number of selected features and maximum accuracy.First,we use the information gain and Fisher Score to sort the features extracted from signals.Then,we employ a multi-objective ranking method to evaluate these features and assign different importance to them.Features with high rankings have a large probability of being selected.Finally,we propose a repair strategy to address the problem of duplicate solutions in multi-objective feature selection,which can improve the diversity of solutions and avoid falling into local traps.Using random forest and K-nearest neighbor classifiers,four English speech emotion datasets are employed to test the proposed algorithm(MBEO)as well as other multi-objective emotion identification techniques.The results illustrate that it performs well in inverted generational distance,hypervolume,Pareto solutions,and execution time,and MBEO is appropriate for high-dimensional English SER.

Exploring Sequential Feature Selection in Deep Bi-LSTM Models for Speech Emotion Recognition

Machine Learning(ML)algorithms play a pivotal role in Speech Emotion Recognition(SER),although they encounter a formidable obstacle in accurately discerning a speaker’s emotional state.The examination of the emotional states of speakers holds significant importance in a range of real-time applications,including but not limited to virtual reality,human-robot interaction,emergency centers,and human behavior assessment.Accurately identifying emotions in the SER process relies on extracting relevant information from audio inputs.Previous studies on SER have predominantly utilized short-time characteristics such as Mel Frequency Cepstral Coefficients(MFCCs)due to their ability to capture the periodic nature of audio signals effectively.Although these traits may improve their ability to perceive and interpret emotional depictions appropriately,MFCCS has some limitations.So this study aims to tackle the aforementioned issue by systematically picking multiple audio cues,enhancing the classifier model’s efficacy in accurately discerning human emotions.The utilized dataset is taken from the EMO-DB database,preprocessing input speech is done using a 2D Convolution Neural Network(CNN)involves applying convolutional operations to spectrograms as they afford a visual representation of the way the audio signal frequency content changes over time.The next step is the spectrogram data normalization which is crucial for Neural Network(NN)training as it aids in faster convergence.Then the five auditory features MFCCs,Chroma,Mel-Spectrogram,Contrast,and Tonnetz are extracted from the spectrogram sequentially.The attitude of feature selection is to retain only dominant features by excluding the irrelevant ones.In this paper,the Sequential Forward Selection(SFS)and Sequential Backward Selection(SBS)techniques were employed for multiple audio cues features selection.Finally,the feature sets composed from the hybrid feature extraction methods are fed into the deep Bidirectional Long Short Term Memory(Bi-LSTM)network to discern emotions.Since the deep Bi-LSTM can hierarchically learn complex features and increases model capacity by achieving more robust temporal modeling,it is more effective than a shallow Bi-LSTM in capturing the intricate tones of emotional content existent in speech signals.The effectiveness and resilience of the proposed SER model were evaluated by experiments,comparing it to state-of-the-art SER techniques.The results indicated that the model achieved accuracy rates of 90.92%,93%,and 92%over the Ryerson Audio-Visual Database of Emotional Speech and Song(RAVDESS),Berlin Database of Emotional Speech(EMO-DB),and The Interactive Emotional Dyadic Motion Capture(IEMOCAP)datasets,respectively.These findings signify a prominent enhancement in the ability to emotional depictions identification in speech,showcasing the potential of the proposed model in advancing the SER field.

Based on STM32 of CNN Speech Keyword Command Recognition System

Speech recognition is a hot topic in the field of artificial intelligence.Generally,speech recognition models can only run on large servers or dedicated chips.This paper presents a keyword speech recognition system based on a neural network and a conventional STM32 chip.To address the limited Flash and ROM resources on the STM32 MCU chip,the deployment of the speech recognition model is optimized to meet the requirements of keyword recognition.Firstly,the audio information obtained through sensors is subjected to MFCC(Mel Frequency Cepstral Coefficient)feature extraction,and the extracted MFCC features are input into a CNN(Convolutional Neural Network)for deep feature extraction.Then,the features are input into a fully connected layer,and finally,the speech keyword is classified and predicted.Deploying the model to the STM32F429,the prediction model achieves an accuracy of 90.58%,a decrease of less than 1%compared to the accuracy of 91.49%running on a computer,with good performance.

Challenges and Limitations in Speech Recognition Technology:A Critical Review of Speech Signal Processing Algorithms,Tools and Systems

Speech recognition systems have become a unique human-computer interaction(HCI)family.Speech is one of the most naturally developed human abilities;speech signal processing opens up a transparent and hand-free computation experience.This paper aims to present a retrospective yet modern approach to the world of speech recognition systems.The development journey of ASR(Automatic Speech Recognition)has seen quite a few milestones and breakthrough technologies that have been highlighted in this paper.A step-by-step rundown of the fundamental stages in developing speech recognition systems has been presented,along with a brief discussion of various modern-day developments and applications in this domain.This review paper aims to summarize and provide a beginning point for those starting in the vast field of speech signal processing.Since speech recognition has a vast potential in various industries like telecommunication,emotion recognition,healthcare,etc.,this review would be helpful to researchers who aim at exploring more applications that society can quickly adopt in future years of evolution.

Automatic depression recognition by intelligent speech signal processing:A systematic survey

Depression has become one of the most common mental illnesses in the world.For better prediction and diagnosis,methods of automatic depression recognition based on speech signal are constantly proposed and updated,with a transition from the early traditional methods based on hand‐crafted features to the application of architectures of deep learning.This paper systematically and precisely outlines the most prominent and up‐to‐date research of automatic depression recognition by intelligent speech signal processing so far.Furthermore,methods for acoustic feature extraction,algorithms for classification and regression,as well as end to end deep models are investigated and analysed.Finally,general trends are summarised and key unresolved issues are identified to be considered in future studies of automatic speech depression recognition.

SPEECH ENHANCEMENT METHOD FOR LPC AUTOREGRESSIVE MODEL AND SYSTEM IMPLEMENTATION OF COMMAND WORD RECOGNITION USED IN NOISY ENVIRONMENT

At present, almost all the systems and products for speech recognition are working in quiet environment and their performances are degraded or even can′t work when they are operated in high noisy environment. In this paper, after analyzing the features of speech and noise, a speech enhancement method for LPC autoregressive model for command words recognition used in noisy environment is proposed, and an experimental system is realized. In different background noisy environments, we conduct experiments about SNR, basic accuracy, noise resistant ability and system environment adaptability with different microphones. The experimental results show that the system has good recognition performance in high noisy environments. The system can resist many kinds of noises and meet the needs of application areas on the whole such as military, traffic, marketplace and factory etc.

“REAL-TIME SPEECH RECOGNITION SYSTEM”HONORED AN INTERNATIONAL AWARD

"Real-Time Speech Recognition System"(RTSRS),developed by Institute of Acous-tics Academia SINICA,was honored by an international award(25,000FF)at hightechnolgy exhibition《Technologie et Comp(?)tivit(?)》held at Grenoble of France onOctober 1988.More than 10,000 exhibits from 12 countries were presented and4 exhibits were honored awards.The technical commitee of the exhibition "Real-Time Speech Recognition System"(RTSRS),developed by Institute of Acous-tics Academia SINICA,was honored by an international award(25,000FF)at hightechnolgy exhibition《Technologie et Compétivité》held at Grenoble of France onOctober 1988.More than 10,000 exhibits from 12 countries were presented and4 exhibits were honored awards.The technical commitee of the exhibitionacknowleged that the RTSRS has special advanced characteristics both in technicalimprovement and in economic usage.(YU Tiecheng)

Multilayer Neural Network Based Speech Emotion Recognition for Smart Assistance

Day by day,biometric-based systems play a vital role in our daily lives.This paper proposed an intelligent assistant intended to identify emotions via voice message.A biometric system has been developed to detect human emotions based on voice recognition and control a few electronic peripherals for alert actions.This proposed smart assistant aims to provide a support to the people through buzzer and light emitting diodes(LED)alert signals and it also keep track of the places like households,hospitals and remote areas,etc.The proposed approach is able to detect seven emotions:worry,surprise,neutral,sadness,happiness,hate and love.The key elements for the implementation of speech emotion recognition are voice processing,and once the emotion is recognized,the machine interface automatically detects the actions by buzzer and LED.The proposed system is trained and tested on various benchmark datasets,i.e.,Ryerson Audio-Visual Database of Emotional Speech and Song(RAVDESS)database,Acoustic-Phonetic Continuous Speech Corpus(TIMIT)database,Emotional Speech database(Emo-DB)database and evaluated based on various parameters,i.e.,accuracy,error rate,and time.While comparing with existing technologies,the proposed algorithm gave a better error rate and less time.Error rate and time is decreased by 19.79%,5.13 s.for the RAVDEES dataset,15.77%,0.01 s for the Emo-DB dataset and 14.88%,3.62 for the TIMIT database.The proposed model shows better accuracy of 81.02%for the RAVDEES dataset,84.23%for the TIMIT dataset and 85.12%for the Emo-DB dataset compared to Gaussian Mixture Modeling(GMM)and Support Vector Machine(SVM)Model.

A Multi-Level Circulant Cross-Modal Transformer for Multimodal Speech Emotion Recognition

Speech emotion recognition,as an important component of humancomputer interaction technology,has received increasing attention.Recent studies have treated emotion recognition of speech signals as a multimodal task,due to its inclusion of the semantic features of two different modalities,i.e.,audio and text.However,existing methods often fail in effectively represent features and capture correlations.This paper presents a multi-level circulant cross-modal Transformer(MLCCT)formultimodal speech emotion recognition.The proposed model can be divided into three steps,feature extraction,interaction and fusion.Self-supervised embedding models are introduced for feature extraction,which give a more powerful representation of the original data than those using spectrograms or audio features such as Mel-frequency cepstral coefficients(MFCCs)and low-level descriptors(LLDs).In particular,MLCCT contains two types of feature interaction processes,where a bidirectional Long Short-term Memory(Bi-LSTM)with circulant interaction mechanism is proposed for low-level features,while a two-stream residual cross-modal Transformer block is appliedwhen high-level features are involved.Finally,we choose self-attention blocks for fusion and a fully connected layer to make predictions.To evaluate the performance of our proposed model,comprehensive experiments are conducted on three widely used benchmark datasets including IEMOCAP,MELD and CMU-MOSEI.The competitive results verify the effectiveness of our approach.

Design of Hierarchical Classifier to Improve Speech Emotion Recognition

Automatic Speech Emotion Recognition(SER)is used to recognize emotion from speech automatically.Speech Emotion recognition is working well in a laboratory environment but real-time emotion recognition has been influenced by the variations in gender,age,the cultural and acoustical background of the speaker.The acoustical resemblance between emotional expressions further increases the complexity of recognition.Many recent research works are concentrated to address these effects individually.Instead of addressing every influencing attribute individually,we would like to design a system,which reduces the effect that arises on any factor.We propose a two-level Hierarchical classifier named Interpreter of responses(IR).Thefirst level of IR has been realized using Support Vector Machine(SVM)and Gaussian Mixer Model(GMM)classifiers.In the second level of IR,a discriminative SVM classifier has been trained and tested with meta information offirst-level classifiers along with the input acoustical feature vector which is used in primary classifiers.To train the system with a corpus of versatile nature,an integrated emotion corpus has been composed using emotion samples of 5 speech corpora,namely;EMO-DB,IITKGP-SESC,SAVEE Corpus,Spanish emotion corpus,CMU's Woogle corpus.The hierarchical classifier has been trained and tested using MFCC and Low-Level Descriptors(LLD).The empirical analysis shows that the proposed classifier outperforms the traditional classifiers.The proposed ensemble design is very generic and can be adapted even when the number and nature of features change.Thefirst-level classifiers GMM or SVM may be replaced with any other learning algorithm.

Joint On-Demand Pruning and Online Distillation in Automatic Speech Recognition Language Model Optimization

Automatic speech recognition(ASR)systems have emerged as indispensable tools across a wide spectrum of applications,ranging from transcription services to voice-activated assistants.To enhance the performance of these systems,it is important to deploy efficient models capable of adapting to diverse deployment conditions.In recent years,on-demand pruning methods have obtained significant attention within the ASR domain due to their adaptability in various deployment scenarios.However,these methods often confront substantial trade-offs,particularly in terms of unstable accuracy when reducing the model size.To address challenges,this study introduces two crucial empirical findings.Firstly,it proposes the incorporation of an online distillation mechanism during on-demand pruning training,which holds the promise of maintaining more consistent accuracy levels.Secondly,it proposes the utilization of the Mogrifier long short-term memory(LSTM)language model(LM),an advanced iteration of the conventional LSTM LM,as an effective alternative for pruning targets within the ASR framework.Through rigorous experimentation on the ASR system,employing the Mogrifier LSTM LM and training it using the suggested joint on-demand pruning and online distillation method,this study provides compelling evidence.The results exhibit that the proposed methods significantly outperform a benchmark model trained solely with on-demand pruning methods.Impressively,the proposed strategic configuration successfully reduces the parameter count by approximately 39%,all the while minimizing trade-offs.

Speech Recognition via CTC-CNN Model

In the speech recognition system,the acoustic model is an important underlying model,and its accuracy directly affects the performance of the entire system.This paper introduces the construction and training process of the acoustic model in detail and studies the Connectionist temporal classification(CTC)algorithm,which plays an important role in the end-to-end framework,established a convolutional neural network(CNN)combined with an acoustic model of Connectionist temporal classification to improve the accuracy of speech recognition.This study uses a sound sensor,ReSpeakerMic Array v2.0.1,to convert the collected speech signals into text or corresponding speech signals to improve communication and reduce noise and hardware interference.The baseline acousticmodel in this study faces challenges such as long training time,high error rate,and a certain degree of overfitting.The model is trained through continuous design and improvement of the relevant parameters of the acousticmodel,and finally the performance is selected according to the evaluation index.Excellentmodel,which reduces the error rate to about 18%,thus improving the accuracy rate.Finally,comparative verificationwas carried out from the selection of acoustic feature parameters,the selection of modeling units,and the speaker’s speech rate,which further verified the excellent performance of the CTCCNN_5+BN+Residual model structure.In terms of experiments,to train and verify the CTC-CNN baseline acoustic model,this study uses THCHS-30 and ST-CMDS speech data sets as training data sets,and after 54 epochs of training,the word error rate of the acoustic model training set is 31%,the word error rate of the test set is stable at about 43%.This experiment also considers the surrounding environmental noise.Under the noise level of 80∼90 dB,the accuracy rate is 88.18%,which is the worst performance among all levels.In contrast,at 40–60 dB,the accuracy was as high as 97.33%due to less noise pollution.

A Robust Conformer-Based Speech Recognition Model for Mandarin Air Traffic Control

This study aims to address the deviation in downstream tasks caused by inaccurate recognition results when applying Automatic Speech Recognition(ASR)technology in the Air Traffic Control(ATC)field.This paper presents a novel cascaded model architecture,namely Conformer-CTC/Attention-T5(CCAT),to build a highly accurate and robust ATC speech recognition model.To tackle the challenges posed by noise and fast speech rate in ATC,the Conformer model is employed to extract robust and discriminative speech representations from raw waveforms.On the decoding side,the Attention mechanism is integrated to facilitate precise alignment between input features and output characters.The Text-To-Text Transfer Transformer(T5)language model is also introduced to handle particular pronunciations and code-mixing issues,providing more accurate and concise textual output for downstream tasks.To enhance the model’s robustness,transfer learning and data augmentation techniques are utilized in the training strategy.The model’s performance is optimized by performing hyperparameter tunings,such as adjusting the number of attention heads,encoder layers,and the weights of the loss function.The experimental results demonstrate the significant contributions of data augmentation,hyperparameter tuning,and error correction models to the overall model performance.On the Our ATC Corpus dataset,the proposed model achieves a Character Error Rate(CER)of 3.44%,representing a 3.64%improvement compared to the baseline model.Moreover,the effectiveness of the proposed model is validated on two publicly available datasets.On the AISHELL-1 dataset,the CCAT model achieves a CER of 3.42%,showcasing a 1.23%improvement over the baseline model.Similarly,on the LibriSpeech dataset,the CCAT model achieves a Word Error Rate(WER)of 5.27%,demonstrating a performance improvement of 7.67%compared to the baseline model.Additionally,this paper proposes an evaluation criterion for assessing the robustness of ATC speech recognition systems.In robustness evaluation experiments based on this criterion,the proposed model demonstrates a performance improvement of 22%compared to the baseline model.

Using Speaker-Specific Emotion Representations in Wav2vec 2.0-Based Modules for Speech Emotion Recognition

Speech emotion recognition is essential for frictionless human-machine interaction,where machines respond to human instructions with context-aware actions.The properties of individuals’voices vary with culture,language,gender,and personality.These variations in speaker-specific properties may hamper the performance of standard representations in downstream tasks such as speech emotion recognition(SER).This study demonstrates the significance of speaker-specific speech characteristics and how considering them can be leveraged to improve the performance of SER models.In the proposed approach,two wav2vec-based modules(a speaker-identification network and an emotion classification network)are trained with the Arcface loss.The speaker-identification network has a single attention block to encode an input audio waveform into a speaker-specific representation.The emotion classification network uses a wav2vec 2.0-backbone as well as four attention blocks to encode the same input audio waveform into an emotion representation.These two representations are then fused into a single vector representation containing emotion and speaker-specific information.Experimental results showed that the use of speaker-specific characteristics improves SER performance.Additionally,combining these with an angular marginal loss such as the Arcface loss improves intra-class compactness while increasing inter-class separability,as demonstrated by the plots of t-distributed stochastic neighbor embeddings(t-SNE).The proposed approach outperforms previous methods using similar training strategies,with a weighted accuracy(WA)of 72.14%and unweighted accuracy(UA)of 72.97%on the Interactive Emotional Dynamic Motion Capture(IEMOCAP)dataset.This demonstrates its effectiveness and potential to enhance human-machine interaction through more accurate emotion recognition in speech.

An Optimal Method for Speech Recognition Based on Neural Network

Natural language processing technologies have become more widely available in recent years,making them more useful in everyday situations.Machine learning systems that employ accessible datasets and corporate work to serve the whole spectrum of problems addressed in computational linguistics have lately yielded a number of promising breakthroughs.These methods were particularly advantageous for regional languages,as they were provided with cut-ting-edge language processing tools as soon as the requisite corporate information was generated.The bulk of modern people are unconcerned about the importance of reading.Reading aloud,on the other hand,is an effective technique for nour-ishing feelings as well as a necessary skill in the learning process.This paper pro-posed a novel approach for speech recognition based on neural networks.The attention mechanism isfirst utilized to determine the speech accuracy andfluency assessments,with the spectrum map as the feature extraction input.To increase phoneme identification accuracy,reading precision,for example,employs a new type of deep speech.It makes use of the exportchapter tool,which provides a corpus,as well as the TensorFlow framework in the experimental setting.The experimentalfindings reveal that the suggested model can more effectively assess spoken speech accuracy and readingfluency than the old model,and its evalua-tion model’s score outcomes are more accurate.

Improved Speech Emotion Recognition Focusing on High-Level Data Representations and Swift Feature Extraction Calculation

The performance of a speech emotion recognition(SER)system is heavily influenced by the efficacy of its feature extraction techniques.The study was designed to advance the field of SER by optimizing feature extraction tech-niques,specifically through the incorporation of high-resolution Mel-spectrograms and the expedited calculation of Mel Frequency Cepstral Coefficients(MFCC).This initiative aimed to refine the system’s accuracy by identifying and mitigating the shortcomings commonly found in current approaches.Ultimately,the primary objective was to elevate both the intricacy and effectiveness of our SER model,with a focus on augmenting its proficiency in the accurate identification of emotions in spoken language.The research employed a dual-strategy approach for feature extraction.Firstly,a rapid computation technique for MFCC was implemented and integrated with a Bi-LSTM layer to optimize the encoding of MFCC features.Secondly,a pretrained ResNet model was utilized in conjunction with feature Stats pooling and dense layers for the effective encoding of Mel-spectrogram attributes.These two sets of features underwent separate processing before being combined in a Convolutional Neural Network(CNN)outfitted with a dense layer,with the aim of enhancing their representational richness.The model was rigorously evaluated using two prominent databases:CMU-MOSEI and RAVDESS.Notable findings include an accuracy rate of 93.2%on the CMU-MOSEI database and 95.3%on the RAVDESS database.Such exceptional performance underscores the efficacy of this innovative approach,which not only meets but also exceeds the accuracy benchmarks established by traditional models in the field of speech emotion recognition.