Yunho Kim

Ulsan National Institute of Science and Technology | UNIST · Mathematics

Appropriate weight initialization settings, along with the ReLU activation function, have become cornerstones of modern deep learning, enabling the training and deployment of highly effective and efficient neural network models across diverse areas of artificial intelligence. The problem of "dying ReLU," where ReLU neurons become inactive and yiel...

Sign language plays a pivotal role in facilitating communication for the deaf community, bridging the gap with the broader society. Nevertheless, mastering sign language poses significant challenges due to the intricate nuances of body movements, hand gestures, and facial expressions. Sign language recognition technology is a pivotal solution aimed...

Maize stands as a cornerstone of Indonesia's agricultural landscape, serving both as a vital food source and an essential fortifier. However, the marketing process of smallholder maize in Indonesia has yet to reach an optimal level of efficiency. This research endeavors to delve into the vertical integration of smallholder maize in the Indonesian a...

The rising prevalence of behavioral and mental health challenges among Indonesian adolescents, affecting nearly one in three, is a pressing concern intensified by constant internet exposure and heightened social comparisons. This study, using data from Indonesia’s National Socioeconomic Survey, examines the likelihood of mental health issues in ado...

Particulate matter forecasting is fundamental for early warning and controlling air pollution, especially PM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2.5</sub> . The increase in this level of concentration will lead to a negative impact on public health. This study develops a hybrid model of CNN...

This research endeavors to conduct a thorough investigation into pediatric health in the Papua and West Papua regions of Indonesia, employing a multifaceted approach that integrates data from various sources including RISKESDAS, SUSENAS, and remote sensing indicators such as NDVI (Normalized Difference Vegetation Index) and PDSI (Palmer Drought Sev...

Reservoir computing, one of the state-of-the-art machine learning architectures, processes time-series data generated by dynamical systems. Nevertheless, we have realized that reservoir computing with the conventional single-reservoir structure suffers from capacity saturation. This leads to performance stagnation in practice. Therefore, we propose...

Air quality conditions are now more severe in the Jakarta area that is among the world’s top eight worst cities according to the 2022 Air Quality Index (AQI) report. In particular, the data from the Meteorological, Climatological, and Geophysical Agency (BMKG) of the Republic of Indonesia, the latest outcomes in air quality conditions in Jakarta an...

This study examines the interplay of air connectivity, sports events, infrastructures, and fiscal support during the period 2017 and 2022 in a designated area called Special Economic Zone in Mandalika, Lombok Island, West Nusa Tenggara to boost tourism development in Indonesia by utilizing big data cognitive analytics. We examine the tourism develo...

2E-MDCVRP is a multi-depot, capacity, two-echelon vehicle routing issue. Satellites make it possible to collect and sort orders. In this paper, a 2E-MDCVRP model will be designed to be applied to e-commerce. In the completion stage, the researcher developed a 2E-MDCVRP solution with two stages of completion. The first stage is finding the best rout...

Amid its massive increase in energy demand, Southeast Asia has pledged to increase its use of renewable energy by up to 23% by 2025. Geospatial technology approaches that integrate statistical data, spatial models, earth observation satellite data, and climate modeling can be used to conduct strategic analyses for understanding the potential and ef...

Phase-field modeling is strongly influenced by the shape of a free energy functional. In the theory of thermodynamics, it is a logarithmic type potential that is legitimate for modeling and simulating binary systems. Nevertheless, a tremendous amount of works have been dedicated to phase-field equations driven by 4-th double-well potentials as a po...

Despite all the expectations for photoacoustic endoscopy (PAE), there are still several technical issues that must be resolved before the technique can be successfully translated into clinics. Among these, electromagnetic interference (EMI) noise, in addition to the limited signal-to-noise ratio (SNR), have hindered the rapid development of related...

We propose a random Fourier sampling scheme to enhance the accuracy of the high frequency pattern estimation for image reconstruction. This method is designed to work in a constrained ℓ1 minimization based on the Fourier-Haar interplay revealing a column-wise maximum coherent structure that we provide. Essential in the scheme is to generate a data-...

While network-based techniques have shown outstanding performance in image denoising in the big data regime requiring massive datasets and expensive computation, mathematical understanding of their working principles is very limited. Not to mention, their relevance to traditional mathematical approaches has not attracted much attention. Therefore,...

There are several theoretically well-posed models for the Allen–Cahn equation under mass conservation. The conservative property is a gift from the additional nonlocal term playing a role of a Lagrange multiplier. However, the same term destroys the boundedness property that the original Allen–Cahn equation presents: The solution is bounded by 1 wi...

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Structural rearrangements govern the various properties of disordered systems and visualization of these dynamical processes can provide critical information on structural deformation and phase transformation of the systems. However, direct imaging of individual atoms or molecules in a disordered state is quite challenging. Here, we prepare a model...

The current work is a continuation of Kim (An unconstrained global optimization framework for real symmetric eigenvalue problems, submitted), where an unconstrained optimization problem was proposed and a first order method was shown to converge to a global minimizer that is an eigenvector corresponding to the smallest eigenvalue with no eigenvalue...

In this work, we interpret real symmetric eigenvalue problems in an unconstrained global optimization framework. More precisely, given two matrices, a symmetric matrix A, and a symmetric positive definite matrix B, we propose and analyze a nonconvex functional F whose local minimizers are, indeed, global minimizers. These minimizers correspond to e...

One important task in image segmentation is to find a region of interest, which is, in general, a solution of a nonlinear and nonconvex problem. The authors of Chan and Esedoglu (Aspects of total variation regularized function approximation. SIAM J. Appl. Math. 2005;65:1817–1837) proposed a convex problem for finding such a region Σ and proved that...

We propose a nonconvex unconstrained minimization problem for eigenvalue problems. In this framework, given a symmetric matrix $A$, it turns out that any nonzero critical point is an eigenvector of $A$ and any local minimizer is a global minimizer, an eigenvector of $A$ corresponding to the smallest eigenvalue. Unlike the conventional way of estima...

Intensity nonuniformity in magnetic resonance (MR) images, represented by a smooth and slowly varying function, is a typical artifact that is a nuisance for many image processing methods. To eliminate the artifact, we have to estimate the nonuniformity as a smooth and slowly varying function and factor it out from the given data. We reformulate the...

This work is devoted to image restoration (denoising and deblurring) by variational models. As in our prior work [Y. Kim and L. A. Vese, Inverse Probl. Imaging 3, No. 1, 43–68 (2009; Zbl 1187.35280)], the image f ˜ to be restored is assumed to be the sum of a cartoon component u (a function of bounded variation) and a texture component v (an oscill...

The presence of noise in High Angular Resolution Diffusion Imaging (HARDI) data of the brain can limit the accuracy with which fiber pathways of the brain can be extracted. In this work, we present a variational model to denoise HARDI data corrupted by Rician noise. We formulate a minimization model composed of a data fidelity term incorporating th...

In this paper, we show additional properties of the limit of a sequence produced by the subspace correction algorithm proposed by Fornasier and Schonlieb [SIAM J. Numer. Anal., 47 (2009), pp. 33973428] for L-2/TV-minimization problems. An important but missing property of such a limiting sequence in that paper is the convergence to a minimizer of t...

A new class of anisotropic diffusion models is proposed for image processing which can be viewed either as a novel kind of regularization of the classical Perona-Malik model or, as advocated by the authors, as a new independent model. The models are diffusive in nature and are characterized by the presence of both forward and backward regimes. In c...

The presence of noise in High Angular Resolution Diffusion Imaging (HARDI) data of the brain can limit the accuracy with which fiber pathways of the brain can be extracted. In this work, we present a variational model to denoise HARDI data corrupted by Rician noise. Numerical experiments are performed on three types of data: 2D synthetic data, 3D d...

We propose a new variational model for image restoration using BV and Sobolev spaces. It is well known that homogeneous Sobolev spaces of negative differentiability can capture oscillatory information very well, however, just one Sobolev space hardly recognizes any difference between texture and noise. By a means of learning a series of Sobolev nor...

In this work, we wish to denoise HARDI (High Angular Resolution Diffusion Imaging) data arising in medical brain imaging. Diffusion imaging is a relatively new and powerful method to measure the three-dimensional profile of water diffusion at each point in the brain. These images can be used to reconstruct fiber directions and pathways in the livin...

In this work we wish to recover an unknown image from a blurry, or noisy-blurry version. We solve this inverse problem by energy minimiza- tion and regularization. We seek a solution of the form u + v, where u is a function of bounded variation (cartoon component), while v is an oscillatory component (texture), modeled by a Sobolev function with ne...

We denoise HARDI (High Angular Resolution Diffusion Imaging) data arising in medical imaging. Diffusion imaging is a relatively new and powerful method to measure the 3D profile of water diffusion at each point. This can be used to reconstruct fiber directions and pathways in the living brain, providing detailed maps of fiber integrity and connecti...

In this work we wish to recover an unknown image from a blurry version. We solve this inverse problem by energy minimization and regularization. We seek a solution of the form u + v, where u is a function of bounded variation (cartoon component), while v is an oscillatory component (texture), modeled by a Sobolev function with negative degree of di...