• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.11 no.4
• /
• pp.280-286
• /
• 1986

A newmethod of fast mask operators for edge detection is proposed, which is based on the matrix factorization. The output of each component in the multi-directional mask operator is obtained adding every image pixels in the mask area weighting by corresponding mask element. Therefore, it is same as the result of matrix-vector multiplication like one dimensional transform, i, e, , trasnform of an image vector surrounded by mask with a transform matrix consisted of all the elements of eack mask row by row. In this paper, for the Sobel and Prewitt operators, we find the transform matrices, add up the number of operations factoring these matrices and compare the performances of the proposed method and the standard method. As a result, the number of operations with the proposed method, for Sobel and prewitt operators, without any extra storage element, are reduced by 42.85% and 50% of the standard operations, respectively and in case of an image having 100x100 pixels, the proposed Sobel operator with 301 extra storage locations can be computed by 35.93% of the standard method.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.7
• /
• pp.205-216
• /
• 2022

One of the many tools used for distributed deep learning training is Kubeflow, which runs on Kubernetes, a container orchestration tool. TensorFlow jobs can be managed using the existing operator provided by Kubeflow. However, when considering the distributed deep learning training jobs based on the parameter server architecture, the scheduling policy used by the existing operator does not consider the task affinity of the distributed training job and does not provide the ability to dynamically allocate or release resources. This can lead to long job completion time and low resource utilization rate. Therefore, in this paper we proposes a new operator that efficiently schedules distributed deep learning training jobs to minimize the job completion time and increase resource utilization rate. We implemented the new operator by modifying the existing operator and conducted experiments to evaluate its performance. The experiment results showed that our scheduling policy improved the average job completion time reduction rate of up to 84% and average CPU utilization increase rate of up to 92%.

• Bulletin of the Korean Chemical Society
• /
• v.33 no.3
• /
• pp.1071-1074
• /
• 2012

• Journal of the Korean Mathematical Society
• /
• v.37 no.2
• /
• pp.309-320
• /
• 2000

• Journal of the Korean Mathematical Society
• /
• v.17 no.1
• /
• pp.123-129
• /
• 1980

• East Asian mathematical journal
• /
• v.14 no.2
• /
• pp.311-316
• /
• 1998

We study some spectral properties of the Laplace-Beltrami operator on compact Riemannian manifolds.

• BMB Reports
• /
• v.37 no.6
• /
• pp.709-714
• /
• 2004

The wild-type repressor CI of temperate mycobacteriophage L1 and the temperature-sensitive (ts) repressor CIts391 of a mutant L1 phage, L1cIts391, have been separately overexpressed in E. coli. Both these repressors were observed to specifically bind with the same cognate operator DNA. The operator-binding activity of CIts391 was shown to differ significantly than that of the CI at 32 to $42^{\circ}C$. While 40-95% operator-binding activity was shown to be retained at 35 to $42^{\circ}C$ in CI, more than 75% operator-binding activity was lost in CIts391 at 35 to $38^{\circ}C$, although the latter showed only 10% less binding compared to that of the former at $32^{\circ}C$. The CIts391 showed almost no binding at $42^{\circ}C$. An in vivo study showed that the CI repressor inhibited the growth of a clear plaque former mutant of the L1 phage more strongly than that of the CIts391 repressor at both 32 and $42^{\circ}C$. The half-life of the CIts391-operator complex was found to be about 8 times less than that of the CI-operator complex at $32^{\circ}C$. Interestingly, the repressor-operator complexes preformed at $0^{\circ}C$ have shown varying degrees of resistance to dissociation at the temperatures which inhibit the formation of these complexes are inhibited. The CI repressor, but not that of CIts391, regains most of the DNA-binding activity on cooling to $32^{\circ}C$ after preincubation at 42 to $52^{\circ}C$. All these data suggest that the 131st proline residue at the C-terminal half of CI, which changed to leucine in the CIts391, plays a crucial role in binding the L1 repressor to the cognate operator DNA, although the helix-turn-helix DNA-binding motif of the L1 repressor is located at its N-terminal end.

• Bulletin of the Korean Chemical Society
• /
• v.33 no.3
• /
• pp.779-789
• /
• 2012

In the kinetic theory of dense fluids the many-particle collision bracket integral is given in terms of a classical collision operator defined in the phase space. To find an algorithm to compute the collision bracket integrals, we revisit the eigenvalue problem of the Liouville operator and re-examine the method previously reported [Chem. Phys. 1977, 20, 93]. Then we apply the notion and concept of the eigenfunctions of the Liouville operator and knowledge acquired in the study of the eigenfunctions to cast collision bracket integrals into more convenient and suitable forms for numerical simulations. One of the alternative forms is given in the form of time correlation function. This form, on a further manipulation, assumes a form reminiscent of the Chapman- Enskog collision bracket integrals, but for dense gases and liquids as well as solids. In the dilute gas limit it would give rise precisely to the Chapman-Enskog collision bracket integrals for two-particle collision. The alternative forms obtained are more readily amenable to numerical simulation methods than the collision bracket integrals expressed in terms of a classical collision operator, which requires solution of classical Lippmann-Schwinger integral equations. This way, the aforementioned kinetic theory of dense fluids is made fully accessible by numerical computation/simulation methods, and the transport coefficients thereof are made computationally as accessible as those in the linear response theory.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.43 no.12 s.354
• /
• pp.8-14
• /
• 2006

This paper proposed noise insensitive 4*4 mid frequency-OCT (MF-DCT) focus measure operator. Proposed operator enhanced low power 8*8 MDCT operator to have 4*4 rotationally same form for Gaussian noise. MF-DCT operator acting like band-pass filter suppresses both low-frequency signal useless for focus measure and high-frequency signal affected by impulsive noise. Also it is proved to be linear because it uses the energy of band-pass filtered signal as focus measure. Experimental result shows its superiority by comparing AUM with traditional operators.

• Proceedings of the IEEK Conference
• /
• 2005.11a
• /
• pp.979-982
• /
• 2005

This paper is to be pre-processing that decides the text recognizability and quality contained in natural image. Differentiated with the existing studies, In this paper, it suggests the application of partially unified color models, Coiflet Wavelet and text extraction algorithm that uses the closed curve edge features of LoG (laplacian of gaussian)operator. The text image included in natural image such as signboard has the same hue, saturation and value, and there is a certain thickness as for their feature. Each color element is restructured into closed area by LoG operator, the 2nd differential operator. The text area is contracted by Hough Transform, logical AND-OR operator of each color model and Minimum-Distance classifier. This paper targets natural image into which text area is added regardless of the size and resolution of the image, and it is confirmed to have more excellent performance than other algorithms with many restrictions.