Yun Kyung Lee

Dr. Yun Kyung Lee

Electronics and Telecommunications Research Institute
orcid

Research Focus & Metrics

Academic Career Length

18Years (2009-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
31 January 2024 in ETRI Journal
Verified
Journal Article
Closed access
BKByung Ok Kang
Byung Ok Kang
HJHyung‐Bae Jeon
Hyung‐Bae Jeon
YLYun Kyung Lee
Yun Kyung Lee

28 November 2022 in Electronics
Verified
Journal Article
Open access
PDF file
JBJohn Lorenzo Bautista
John Lorenzo Bautista
YLYun Kyung Lee
Yun Kyung Lee
HSHyun Soon Shin
Hyun Soon Shin
In this paper, an automatic speech emotion recognition (SER) task of classifying eight different emotions was experimented using parallel based networks trained using the Ryeson Audio-Visual Dataset of Speech and Song (RAVDESS) dataset. A combination of a CNN-based network and attention-based networks, running in parallel, was used to model both spatial features and temporal feature representations. Multiple Augmentation techniques using Additive White Gaussian Noise (AWGN), SpecAugment, Room Impulse Response (RIR), and Tanh Distortion techniques were used to augment the training data to further generalize the model representation. Raw audio data were transformed into Mel-Spectrograms as the model’s input. Using CNN’s proven capability in image classification and spatial feature representations, the spectrograms were treated as an image with the height and width represented by the spectrogram’s time and frequency scales. Temporal feature representations were represented by attention-based models Transformer, and BLSTM-Attention modules. Proposed architectures of the parallel CNN-based networks running along with Transformer and BLSTM-Attention modules were compared with standalone CNN architectures and attention-based networks, as well as with hybrid architectures with CNN layers wrapped in time-distributed wrappers stacked on attention-based networks. In these experiments, the highest accuracy of 89.33% for a Parallel CNN-Transformer network and 85.67% for a Parallel CNN-BLSTM-Attention Network were achieved on a 10% hold-out test set from the dataset. These networks showed promising results based on their accuracies, while keeping significantly less training parameters compared with non-parallel hybrid models.

17 April 2009 in IEEE Transactions on Consumer Electronics
Verified
Journal Article
Closed access
YLYun Kyung Lee
Yun Kyung Lee
OKOh-Wook Kwon
Oh-Wook Kwon
This paper addresses the problem of single channel speech separation for robust speech recognition in noisy environments. We propose to apply shape analysis techniques in image analysis to a computational auditory scene analysis (CASA)-based speech separation system. The CASA-based system extracts the desired speech signals from a mixture of speech signals and other sound sources. In the proposed method, we complement the missing speech signals by applying the shape analysis techniques such as labeling and distance function. In the speech separation experiment using various artificial mixtures of speech and various noise sources, the proposed system increases the signal-to-noise ratio by 6.6 dB. In the speech recognition experiment using the Interspeech speech database, it improves recognition accuracy by 22% for recognition of speech added with speech-shaped stationary noise.

16 March 2021 in Applied Sciences
Verified
Journal Article
Open access
PDF file
YLYun Kyung Lee
Yun Kyung Lee
JPJeon Gue Park
Jeon Gue Park
This paper addresses an automatic proficiency evaluation and speech recognition for second language (L2) speech. The proposed method recognizes the speech uttered by the L2 speaker, measures a variety of fluency scores, and evaluates the proficiency of the speaker’s spoken English. Stress and rhythm scores are one of the important factors used to evaluate fluency in spoken English and are computed by comparing the stress patterns and the rhythm distributions to those of native speakers. In order to compute the stress and rhythm scores even when the phonemic sequence of the L2 speaker’s English sentence is different from the native speaker’s one, we align the phonemic sequences based on a dynamic time-warping approach. We also improve the performance of the speech recognition system for non-native speakers and compute fluency features more accurately by augmenting the non-native training dataset and training an acoustic model with the augmented dataset. In this work, we augment the non-native speech by converting some speech signal characteristics (style) while preserving its linguistic information. The proposed variational autoencoder (VAE)-based speech conversion network trains the conversion model by decomposing the spectral features of the speech into a speaker-invariant content factor and a speaker-specific style factor to estimate diverse and robust speech styles. Experimental results show that the proposed method effectively measures the fluency scores and generates diverse output signals. Also, in the proficiency evaluation and speech recognition tests, the proposed method improves the proficiency score performance and speech recognition accuracy for all proficiency areas compared to a method employing conventional acoustic models.

1 October 2014 in ETRI Journal
Verified
Journal Article
Open access
PDF file
JPJeon Gue Park
Jeon Gue Park
YLYun Kyung Lee
Yun Kyung Lee
OKOh-Wook Kwon
Oh-Wook Kwon
We propose a novel phase-based method for single-channel speech enhancement to extract and enhance the desired signals in noisy environments by utilizing the phase information. In the method, a phase-dependent a priori signal-to-noise ratio (SNR) is estimated in the log-mel spectral domain to utilize both the magnitude and phase information of input speech signals. The phase-dependent estimator is incorporated into the conventional magnitude-based decision-directed approach that recursively computes the a priori SNR from noisy speech. Additionally, we reduce the performance degradation owing to the one-frame delay of the estimated phase-dependent a priori SNR by using a minimum mean square error (MMSE)-based and maximum a posteriori (MAP)-based estimator. In our speech enhancement experiments, the proposed phase-dependent a priori SNR estimator is shown to improve the output SNR by 2.6 dB for both the MMSE-based and MAP-based estimator cases as compared to a conventional magnitude-based estimator.

10 February 2021 in IEEE Access
Verified
Journal Article
Open access
YLYun Kyung Lee
Yun Kyung Lee
HKHyun Woo Kim
Hyun Woo Kim
JPJeon Gue Park
Jeon Gue Park
We address a nonparallel data-driven many-to-many speech modeling and multimodal style conversion method. In this work, we train a speech conversion model for multiple domains rather than a specific source and target domain pair, and we generate diverse output speech signals from a given source domain speech by transferring some speech style-related characteristics while preserving its linguistic content information. The proposed method comprises a variational autoencoder (VAE)-based many-to-many speech conversion network with a Wasserstein generative adversarial network (WGAN) and a skip-connected autoencoder-based self-supervised learning network. The proposed conversion network trains the models by decomposing the spectral features of the input speech signal into a content factor that represents domain-invariant information and a style factor that represents domain-related information to automatically estimate the various speech styles of each domain, and the network converts the input speech signal to another domain using the computed content factor with the target style factor we want to change. Diverse and multimodal outputs can be generated by sampling different style factors. We also train models in a stable manner and improve the quality of generated outputs by sharing the discriminator of the VAE-based speech conversion network and that of the self-supervised learning network. We apply the proposed method to speaker conversion and perform the perceptual evaluations. Experimental results revealed that the proposed method obtained high accuracy of converted spectra, significantly improved the sound quality and speaker similarity of the converted speech, and contributed to stable model training.