• The Journal of the Acoustical Society of Korea
• /
• v.43 no.2
• /
• pp.253-259
• /
• 2024

This paper proposes a new speech enhancement model based on a complex nested U-Net with a dual-branch decoder. The proposed model consists of a complex nested U-Net to simultaneously estimate the magnitude and phase components of the speech signal, and the decoder has a dual-branch decoder structure that performs spectral mapping and time-frequency masking in each branch. At this time, compared to the single-branch decoder structure, the dual-branch decoder structure allows noise to be effectively removed while minimizing the loss of speech information. The experiment was conducted on the VoiceBank + DEMAND database, commonly used for speech enhancement model training, and was evaluated through various objective evaluation metrics. As a result of the experiment, the complex nested U-Net-based speech enhancement model using a dual-branch decoder increased the Perceptual Evaluation of Speech Quality (PESQ) score by about 0.13 compared to the baseline, and showed a higher objective evaluation score than recently proposed speech enhancement models.

• Journal of IKEEE
• /
• v.28 no.1
• /
• pp.6-13
• /
• 2024

This paper proposes an approach to detect abnormal data in automotive controller area network (CAN) using an unsupervised learning model, i.e. autoencoder and Gaussian kernel density estimation function. The proposed autoencoder model is trained with only message ID of CAN data frames. Afterwards, by employing the Gaussian kernel density estimation function, it effectively detects abnormal data based on the trained model characterized by the optimally determined number of frames and a loss threshold. It was verified and evaluated using four types of attack data, i.e. DoS attacks, gear spoofing attacks, RPM spoofing attacks, and fuzzy attacks. Compared with conventional unsupervised learning-based models, it has achieved over 99% detection performance across all evaluation metrics.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.1
• /
• pp.301-307
• /
• 2024

This study introduces a novel cascade fusion architecture aimed at enhancing thermal images across various scale conditions. The processing of thermal images at multiple scales has been challenging due to the limitations of existing methods that are designed for specific scales. To overcome these limitations, this paper proposes a unified framework that utilizes cascade feature fusion to effectively learn multi-scale representations. Confidence maps from different image scales are fused in a cascaded manner, enabling scale-invariant learning. The architecture comprises end-to-end trained convolutional neural networks to enhance image quality by reinforcing mutual scale dependencies. Experimental results indicate that the proposed technique outperforms existing methods in multi-scale thermal image enhancement. Performance evaluation results are provided, demonstrating consistent improvements in image quality metrics. The cascade fusion design facilitates robust generalization across scales and efficient learning of cross-scale representations.

• International Journal of Computer Science & Network Security
• /
• v.24 no.2
• /
• pp.178-188
• /
• 2024

In the near future, it is expected that there will be billions of connected devices using fifth generation (5G) network services. The recently available base stations (BSs) need to mitigate their loads without changing and at the least monetary cost. The available spectrum resources are limited and need to be exploited in an efficient way to meet the ever-increasing demand for services. Device to Device communication (D2D) technology will likely help satisfy the rapidly increasing capacity and also effectively offload traffic from the BS by distributing the transmission between D2D users from one side and the cellular users and the BS from the other side. In this paper, we propose to apply D2D overlay communication with cognitive radio capability in 5G networks to exploit unused spectrum resources taking into account the dynamic spectrum access. The performance metrics; throughput and delay are formulated and analyzed for CSMA-based medium access control (MAC) protocol that utilizes a common control channel for device users to negotiate the data channel and address the contention between those users. Device users can exploit the cognitive radio to access the data channels concurrently in the common interference area. Estimating the achievable throughput and delay in D2D communication in 5G networks is not exploited in previous studies using cognitive radio with CSMA-based MAC protocol to address the contention. From performance analysis, applying cognitive radio capability in D2D communication and allocating a common control channel for device users effectively improve the total aggregated network throughput by more than 60% compared to the individual D2D throughput without adding harmful interference to cellular network users. This approach can also reduce the delay.

• International Journal of Computer Science & Network Security
• /
• v.24 no.3
• /
• pp.59-70
• /
• 2024

Background: The COVID-19 pandemic (the form of coronaviruses) developed at the end of 2019 and spread rapidly to almost every corner of the world. It has infected around 25,334,339 of the world population by the end of September 1, 2020 [1] . It has been spreading ever since, and the peak specific to every country has been rising and falling and does not seem to be over yet. Currently, the conventional RT-PCR testing is required to detect COVID-19, but the alternative method for data archiving purposes is certainly another choice for public departments to make. Researchers are trying to use medical images such as X-ray and Computed Tomography (CT) to easily diagnose the virus with the aid of Artificial Intelligence (AI)-based software. Method: This review paper provides an investigation of a newly emerging machine-learning method used to detect COVID-19 from X-ray images instead of using other methods of tests performed by medical experts. The facilities of computer vision enable us to develop an automated model that has clinical abilities of early detection of the disease. We have explored the researchers' focus on the modalities, images of datasets for use by the machine learning methods, and output metrics used to test the research in this field. Finally, the paper concludes by referring to the key problems posed by identifying COVID-19 using machine learning and future work studies. Result: This review's findings can be useful for public and private sectors to utilize the X-ray images and deployment of resources before the pandemic can reach its peaks, enabling the healthcare system with cushion time to bear the impact of the unfavorable circumstances of the pandemic is sure to cause

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.2
• /
• pp.85-95
• /
• 2009

In the binaural synthesis environments, wide varieties of the head-related transfer functions (HRTFs) that have measured with a various direction would be desirable to obtain the accurate and various spatial sound images. To reduce the size' of HRTFs, interpolation has been often employed, where the HRTF for any direction is obtained by a limited number of the representative HRTFs. In this paper, we study on the distortion measures for interpolation, which has an important role in interpolation. With lhe various objective distortion metrics, the differences between the interpolated and the measured HRTFs were computed. These were then compared and analyzed with the results from the listening tests. From the results, the objective distortion measures were selected, that reflected the perceptual differences in spatial sound image. This measure was employed in a practical interpolation technique. We applied the proposed method to four kinds of an HRTF set, measured from three human heads and one mannequin. As a result, the Mel-frequency cepstral distortion was shown to be a good predictor for the differences in spatial sound location, when three HRTF measured from human, and the time-domain signal to distortion ratio revealed good prediction results for the entire four HRTF sets.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.1
• /
• pp.325-335
• /
• 2024

In this study, I investigated cases of operating Sustainable Development Goals (SDGs) sustainable design courses based on interdisciplinary education for diverse design major students in the fashion design department. Through literature review, we examined the necessity of this course operation and analyzed the course through class design, execution, and operational results. Sustainable design courses were organized for 2nd to 4th-year students, promoting integrated learning for fashion design and various design majors to enhance interdisciplinary skills based on the in-depth study of SDGs issues. The educational content in the classes focused on the sustainable development goals achieved through upcycling design of waste PET bottle fibers developed by local industries, aiming to pursue sustainable values of designers through problem discovery and resolution. Students developed various upcycled products, evaluated metrics, and assessed satisfaction levels. Through this process, students gained an understanding of the practical value of SDGs, recognized the importance of sustainable development through design approaches for solving local issues, and acknowledged the significance of interdisciplinary education with various design majors.

• Geomechanics and Engineering
• /
• v.36 no.5
• /
• pp.465-474
• /
• 2024

Uniaxial compressive strength (UCS) is a critical geomechanical parameter that plays a significant role in the evaluation of rocks. The practice of indirectly estimating said characteristics is widespread due to the challenges associated with obtaining high-quality core samples. The primary aim of this study is to investigate the feasibility of utilizing the gene expression programming (GEP) technique for the purpose of forecasting the UCS for various rock categories, including Schist, Granite, Claystone, Travertine, Sandstone, Slate, Limestone, Marl, and Dolomite, which were sourced from a wide range of quarry sites. The present study utilized a total of 170 datasets, comprising Schmidt hammer (SH), porosity (n), point load index (Is(50)), and P-wave velocity (Vp), as the effective parameters in the model to determine their impact on the UCS. The UCS parameter was computed through the utilization of the GEP model, resulting in the generation of an equation. Subsequently, the efficacy of the GEP model and the resultant equation were assessed using various statistical evaluation metrics to determine their predictive capabilities. The outcomes indicate the prospective capacity of the GEP model and the resultant equation in forecasting the unconfined compressive strength (UCS). The significance of this study lies in its ability to enable geotechnical engineers to make estimations of the UCS of rocks, without the requirement of conducting expensive and time-consuming experimental tests. In particular, a user-friendly program was developed based on the GEP model to enable rapid and very accurate calculation of rock's UCS, doing away with the necessity for costly and time-consuming laboratory experiments.

• Journal of Digital Policy
• /
• v.3 no.1
• /
• pp.27-32
• /
• 2024

In this paper, proposes a super resolution method that enhances the quality of results by refining texture features, contrasting each, and utilizing the results as weights. For the improvement of quality, a precise and clear restoration result in details such as boundary areas is crucial in super resolution, along with minimizing unnecessary artifacts like noise. The proposed method constructs a residual block structure with multiple paths and skip-connections for feature estimation in conventional Convolutional Neural Network (CNN)-based super resolution methods to enhance quality. Additional learning is performed for sharpened and blurred image results for further texture analysis. By contrasting each super resolution result and allocating weights through this process, the proposed method achieves improved quality in detailed and smoothed areas of the image. The experimental results of the proposed method, evaluated using the PSNR and SSIM values as quality metrics, show higher results compared to existing algorithms, confirming the enhancement in quality.

• Convergence Security Journal
• /
• v.23 no.4
• /
• pp.61-70
• /
• 2023

Clustering is an unsupervised learning method that involves grouping data based on features such as distance metrics, using data without known labels or ground truth values. This method has the advantage of being applicable to various types of data, including images, text, and audio, without the need for labeling. Traditional clustering techniques involve applying dimensionality reduction methods or extracting specific features to perform clustering. However, with the advancement of deep learning models, research on deep clustering techniques using techniques such as autoencoders and generative adversarial networks, which represent input data as latent vectors, has emerged. In this study, we propose a deep clustering technique based on deep learning. In this approach, we use an autoencoder to transform the input data into latent vectors, and then construct a vector space according to the cluster structure and perform k-means clustering. We conducted experiments using the MNIST and Fashion-MNIST datasets in the PyTorch machine learning library as the experimental environment. The model used is a convolutional neural network-based autoencoder model. The experimental results show an accuracy of 89.42% for MNIST and 56.64% for Fashion-MNIST when k is set to 10.