• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.1
• /
• pp.15-23
• /
• 2022

Among the learning methods for Continuous Learning environments, "Learning without Forgetting" has fixed regularization strengths, which can lead to poor performance in environments where various data are received. We suggest a way to set weights variable by identifying the features of the data we want to learn. We applied weights adaptively using correlation and complexity. Scenarios with various data are used for evaluation and experiments showed accuracy increases by up to 5% in the new task and up to 11% in the previous task. In addition, it was found that the adaptive weight value obtained by the algorithm proposed in this paper, approached the optimal weight value calculated manually by repeated experiments for each experimental scenario. The correlation coefficient value is 0.739, and overall average task accuracy increased. It can be seen that the method of this paper sets an appropriate lambda value every time a new task is learned, and derives the optimal result value in various scenarios.

• The Journal of Korea Robotics Society
• /
• v.19 no.1
• /
• pp.98-105
• /
• 2024

This paper proposes a new approach with a distance-based regularization to the entropy applied to the NBV (Next-Best-View) selection with NeRF (Neural Radiance Fields). 3D reconstruction requires images from various viewpoints, and selecting where to capture these images is a highly complex problem. In a recent work, image acquisition was derived using NeRF's ray-based uncertainty. While this work was effective for evaluating candidate viewpoints at fixed distances from a camera to an object, it is limited when dealing with a range of candidate viewpoints at various distances, because it tends to favor selecting viewpoints at closer distances. Acquiring images from nearby viewpoints is beneficial for capturing surface details. However, with the limited number of images, its image selection is less overlapped and less frequently observed, so its reconstructed result is sensitive to noise and contains undesired artifacts. We propose a method that incorporates distance-based regularization into entropy, allowing us to acquire images at distances conducive to capturing both surface details without undesired noise and artifacts. Our experiments with synthetic images demonstrated that NeRF models with the proposed distance and entropy-based criteria achieved around 50 percent fewer reconstruction errors than the recent work.

• Journal of the Korean Chemical Society
• /
• v.40 no.1
• /
• pp.20-27
• /
• 1996

A stable and accurate inverse method for extracting potential from spectroscopic data studied. The method is based on the Tikhonov regularization method to overcome the possible instability of nonlinear inverse problems using a priori smooth properties of the potential energy surface. The merit of this method is to treat the potential as continuous functions of the intermolecular coordinates instead of the conventional parameter fitting of restricted potential forms. Numerical study for the Ar-Ar show that from spectroscopic data the exact potential has been recovered whole region and the discrepancies by the dispersion force observed at the large distance between the exact and Morse potential or from RKR method can be eliminated by this method.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.557-557
• /
• 2000

Tracking of the moving object, which is realized by the computer vision, is used for military and industrial fields. It is the application technique with imply complicated processing for understanding the input images. But, in these days, the most moving object tracking algorithms have many difficult problems. A typical problem is the increase of calculation time depending on target number. For this reason, there are many studies to solve real time processing problems and errors for background environmental change. In this paper, we used optical flow which is one of moving object tracking algorithms. It represents vector of the moving object. Optical flow estimation based on the regularization method depends on iteration method but it is very sensitive the noise. We proposed a new method using the Combinatorial Hough Transform (CHT) and Voting Accumulation in order to find optimal constraint lines. Also, we used the logical operation in order to release the operation time. The proposed method can easily and accurately extract the optical flow of moving object area and the moving information. We have simulated the proposed method using the test images. This images are included the noise. Experimental results show that the proposed method get better flow and estimate accurately the moving information.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2006.04a
• /
• pp.299-306
• /
• 2006

Using a level set method and topological derivatives, a topological shape optimization method that is independent of an initial design is developed for linearly elastic structures. In the level set method, the initial domain is kept fixed and its boundary is represented by an implicit moving boundary embedded in the level set function, which facilitates to handle complicated topological shape changes. The 'Hamilton-Jacobi (H-J)' equation and computationally robust numerical technique of 'up-wind scheme' lead the initial implicit boundary to an optimal one according to the normal velocity field while minimizing the objective function of compliance and satisfying the constraint of allowable volume. Based on the asymptotic regularization concept, the topological derivative is considered as the limit of shape derivative as the radius of hole approaches to zero. The required velocity field to update the H -J equation is determined from the descent direction of Lagrangian derived from optimality conditions. It turns out that the initial holes is not required to get the optimal result since the developed method can create holes whenever and wherever necessary using indicators obtained from the topological derivatives. It is demonstrated that the proper choice of control parameters for nucleation is crucial for efficient optimization process.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.9
• /
• pp.1985-1997
• /
• 1997

This paper proposed a regularized iterative image restoration by using method of conjugate gradient. Compared with conventional iterative methods, method of conjugate gradient has a merit to converte toward a solution as a super-linear convergence speed. But because of those properties, there are several artifacts like ringing effects and the partial magnification of the noise in the course of restoring the images that are degraded by a defocusing blur and additive noise. So, we proposed the regularized method of conjugate gradient applying constraints. By applying the projectiong constraint and regularization parameter into that method, it is possible to suppress the magnification of the additive noise. As a experimental results, we showed the superior convergence ratio of the proposed mehtod compared with conventional iterative regularized methods.

• Journal of Biomedical Engineering Research
• /
• v.35 no.5
• /
• pp.142-150
• /
• 2014

The key to model-based iterative reconstruction (MBIR) algorithms for transmission computed tomography lies in the ability to accurately model the data formation process from the emitted photons produced in the transmission source to the measured photons at the detector. Therefore, accurately modeling the system matrix that accounts for the data formation process is a prerequisite for MBIR-based algorithms. In this work we compared quantitative performance of the three representative ray-driven methods for calculating the system matrix; the ray-tracing method (RTM), the distance-driven method (DDM), and the strip-area based method (SAM). We implemented the ordered-subsets separable surrogates (OS-SPS) algorithm using the three different models and performed simulation studies using a digital phantom. Our experimental results show that, in spite of the more advanced features in the SAM and DDM, the traditional RTM implemented in the OS-SPS algorithm with an edge-preserving regularizer out-performs the SAM and DDM in restoring complex edges in the underlying object. The performance of the RTM in smooth regions was also comparable to that of the SAM or DDM.

• Current Optics and Photonics
• /
• v.5 no.3
• /
• pp.322-328
• /
• 2021

In this paper, a pretreatment method for a matrix in convex optimization is proposed to optimize the spectral reconstruction process of a disordered dispersion spectrometer. Unlike the reconstruction process of traditional spectrometers using Fourier transforms, the reconstruction process of disordered dispersion spectrometers involves solving a large-scale matrix equation. However, since the matrices in the matrix equation are obtained through measurement, they contain uncertainties due to out of band signals, background noise, rounding errors, temperature variations and so on. It is difficult to solve such a matrix equation by using ordinary nonstationary iterative methods, owing to instability problems. Although the smoothing Tikhonov regularization approach has the ability to approximatively solve the matrix equation and reconstruct most simple spectral shapes, it still suffers the limitations of reconstructing complex and irregular spectral shapes that are commonly used to distinguish different elements of detected targets with mixed substances by characteristic spectral peaks. Therefore, we propose a special pretreatment method for a matrix in convex optimization, which has been proved to be useful for reducing the condition number of matrices in the equation. In comparison with the reconstructed spectra gotten by the previous ordinary iterative method, the spectra obtained by the pretreatment method show obvious accuracy.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.54 no.4
• /
• pp.83-90
• /
• 2017

Electrical impedance tomography is a relatively new nondestructive imaging modality in which the internal conductivity distribution is reconstructed based on the injected currents and measured voltages through electrodes placed on the surface of a domain. In this paper, a fast iterative Gauss-Newton method is proposed to increase the spatial resolution as well as reduce the inverse computational time in the inverse problem, which could be applied to online binary mixture flow applications. To evaluate the reconstruction performance of the proposed method, numerical experiments have been carried out and the results are analyzed.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.6
• /
• pp.2709-2729
• /
• 2016

Representation based classification, kernel method and sparse representation have received much attention in the field of face recognition. In this paper, we proposed an improved kernel principal component analysis method based on sparse representation to improve the accuracy and robustness for face recognition. First, the distances between the test sample and all training samples in kernel space are estimated based on collaborative representation. Second, S training samples with the smallest distances are selected, and Kernel Principal Component Analysis (KPCA) is used to extract the features that are exploited for classification. The proposed method implements the sparse representation under ℓ₂ regularization and performs feature extraction twice to improve the robustness. Also, we investigate the relationship between the accuracy and the sparseness coefficient, the relationship between the accuracy and the dimensionality respectively. The comparative experiments are conducted on the ORL, the GT and the UMIST face database. The experimental results show that the proposed method is more effective and robust than several state-of-the-art methods including Sparse Representation based Classification (SRC), Collaborative Representation based Classification (CRC), KCRC and Two Phase Test samples Sparse Representation (TPTSR).