June-Woo Kim

Prof. Dr. June-Woo Kim

orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
25 February 2025 in IEEE Journal of Biomedical and Health Informatics
Verified
Journal Article
Closed access
JKJune-Woo Kim
June-Woo Kim
MTMiika Toikkanen
Miika Toikkanen
AJAmin Jalali
Amin Jalali
MKMinseok Kim
Minseok Kim
HHHye-Ji Han
Hye-Ji Han
KKKyunghoon Kim
Kyunghoon Kim
Despite considerable advancements in deep learning, optimizing respiratory sound classification (RSC) models remains challenging. This is partly due to the bias from inconsistent respiratory sound recording processes and imbalanced representation of demographics, which leads to poor performance when a model trained with the dataset is applied to real-world use cases. RSC datasets usually include various metadata attributes describing certain aspects of the data, such as environmental and demographic factors. To address the issues caused by bias, we take advantage of the metadata provided by RSC datasets and explore approaches for metadata-guided domain adaptation. We thoroughly evaluate the effect of various metadata attributes and their combinations on a simple metadata-guided approach, but also introduce a more advanced method that adaptively rescales the suitable metadata combinations to improve domain adaptation during training. The findings indicate a robust reduction in domain dependency and improvement in detection accuracy on both ICBHI and our own dataset. Specifically, the implementation of our proposed methods led to an improved score of 84.97%, which signifies a substantial enhancement of 7.37% compared to the baseline model.

24 September 2021 in IEEE Access
Verified
Journal Article
Open access
JKJune-Woo Kim
June-Woo Kim
HYHyekyung Yoon
Hyekyung Yoon
HJHo-Young Jung
Ho-Young Jung
We address a low-performance problem of the elderly in automatic speech recognition (ASR) through feature adaptation agnostic to the ASR. Most of the datasets for speech recognition models consist of datasets collected from adult speakers. Consequently, the majority of commercial speech recognition systems typically tend to perform well on adult speakers. In other words, the limited diversity of speakers in the training datasets yields unreliable performance for minority (e.g., elderly) speakers due to the infeasible acquisition of training data. In response, this paper suggests a neural network-based voice conversion framework to enhance speech recognition of the minority. To this end, we propose a voice translation model including an unsupervised phonology clustering to extract linguistic information to fit the minority’s speech to a current acoustic model frame. Our proposal is a spectral feature adaptation method that can be placed in front of any commercial or open ASR system, avoiding directly modifying the speech recognizer. The experimental results and analysis demonstrate the effectiveness of our proposed method through improvement in elderly speech recognition accuracy.

15 February 2022 in Sensors
Verified
Journal Article
Open access
PDF file
JKJune-Woo Kim
June-Woo Kim
HYHyekyung Yoon
Hyekyung Yoon
HJHo-Young Jung
Ho-Young Jung
Successful applications of deep learning technologies in the natural language processing domain have improved text-based intent classifications. However, in practical spoken dialogue applications, the users’ articulation styles and background noises cause automatic speech recognition (ASR) errors, and these may lead language models to misclassify users’ intents. To overcome the limited performance of the intent classification task in the spoken dialogue system, we propose a novel approach that jointly uses both recognized text obtained by the ASR model and a given labeled text. In the evaluation phase, only the fine-tuned recognized language model (RLM) is used. The experimental results show that the proposed scheme is effective at classifying intents in the spoken dialogue system containing ASR errors.

4 February 2026 in Communications Medicine
Verified
Journal Article
Open access
PDF file
JKJune-Woo Kim
June-Woo Kim
HYHaram Yoon
Haram Yoon
BKBung-Nyun Kim
Bung-Nyun Kim
SLSang-Yeol Lee
Sang-Yeol Lee
DKDae-Jin Kim
Dae-Jin Kim
CYChan-Mo Yang
Chan-Mo Yang
Background In adolescents, identifying objective biomarkers for treatment response is crucial for the development of effective interventions. Voice-based biomarkers have recently shown potential to capture treatment-related changes in Major Depressive Disorder (MDD). While prior studies have been cross-sectional experiments with single speech sample, this study addresses a critical gap by evaluating intra-patient changes in speech over treatment period, providing insight into how these voice biomarkers evolve within individuals. Methods We collected pre- and post-treatment voice samples from 48 adolescent MDD patients. We hypothesized that deep learning models could detect clinically meaningful changes in depressive states during treatment. Therefore, we compared machine learning and deep learning models for depressive classification. Additionally, we introduced the Dual Voice-based Depressive State Analysis (DVDSA) method to categorize intra-patient depressive state changes as recovery, worsening, or unchanged, highlighting the deep learning models’ ability to detect these variations. Results Among the acoustic features, only the fundamental frequency exhibits significant changes between pre- and post-treatment states after Holm-Bonferroni correction. Machine learning models demonstrate limited performance in distinguishing treatment states, with the best F1-score reaching 65.83%. In contrast, deep learning model, particularly WavLM, achieves remarkably higher performance in binary classification, with an F1-score of 78.05%. The WavLM maintains robust performance, when applied to the DVDSA method, achieves an F1-score of 70.58%. Conclusions These findings suggest that machine learning models and individual acoustic features may not sufficiently capture treatment-related changes in MDD patients. This study underscores the value of deep learning models using the DVDSA method, addressing the limitations of pre- and post-treatment classification and highlighting their potential to advance personalized treatment strategies for adolescent MDD.

26 January 2023 in Mathematics
Verified
Journal Article
Open access
PDF file
JKJune-Woo Kim
June-Woo Kim
HCHoon Chung
Hoon Chung
HJHo-Young Jung
Ho-Young Jung
Unsupervised learning-based approaches for training speech vector representations (SVR) have recently been widely applied. While pretrained SVR models excel in relatively clean automatic speech recognition (ASR) tasks, such as those recorded in laboratory environments, they are still insufficient for practical applications with various types of noise, intonation, and dialects. To cope with this problem, we present a novel unsupervised SVR learning method for practical end-to-end ASR models. Our approach involves designing a speech feature masking method to stabilize SVR model learning and improve the performance of the ASR model in a downstream task. By introducing a noise masking strategy into diverse combinations of the time and frequency regions of the spectrogram, the SVR model becomes a robust representation extractor for the ASR model in practical scenarios. In pretraining experiments, we train the SVR model using approximately 18,000 h of Korean speech datasets that included diverse speakers and were recorded in environments with various amounts of noise. The weights of the pretrained SVR extractor are then frozen, and the extracted speech representations are used for ASR model training in a downstream task. The experimental results show that the ASR model using our proposed SVR extractor significantly outperforms conventional methods.

5 August 2023 in Mathematics
Verified
Journal Article
Open access
PDF file
JKJune-Woo Kim
June-Woo Kim
HCHoon Chung
Hoon Chung
HJHo-Young Jung
Ho-Young Jung
Recent advanced systems in the speech recognition domain use large Transformer neural networks that have been pretrained on massive speech data. General methods in the deep learning area have been frequently shared across various domains, and the Transformer model can also be used effectively across speech and image. In this paper, we introduce a novel masking method for self-supervised speech representation learning with salt-and-pepper (S&P) mask which is commonly used in computer vision. The proposed scheme includes consecutive quadrilateral-shaped S&P patches randomly contaminating the input speech spectrum. Furthermore, we modify the standard S&P mask to make it appropriate for the speech domain. In order to validate the effect of the proposed spectral S&P patch masking for the self-supervised representation learning approach, we conduct the pretraining and downstream experiments with two languages, English and Korean. To this end, we pretrain the speech representation model using each dataset and evaluate the pretrained models for feature extraction and fine-tuning performance on varying downstream tasks, respectively. The experimental outcomes clearly illustrate that the proposed spectral S&P patch masking is effective for various downstream tasks when combined with the conventional masking methods.