• Honam Mathematical Journal
• /
• v.9 no.1
• /
• pp.135-147
• /
• 1987

It is well known that there are two ways to define Chern classes of complex vector bundles. One gives the definition of Chern classes by the five axioms ([2]. [3], [4]). and an other defines Chern classes with the associated projective space bundle of a given bundle ([1]. [5]). The purpose of this paper is to describe the latter way in detail and to give new proofs of that our Chern classes satisfy the five axioms with respect to Chern classes (for example Theorem 5).

• Bulletin of the Korean Mathematical Society
• /
• v.35 no.1
• /
• pp.119-125
• /
• 1998

We determine the possible types of Chern classes of rank 2 stable bundles (mod Picard groups) on the Enriques surfaces and their covering K3 surfaces.

• Journal of the Korean Mathematical Society
• /
• v.47 no.6
• /
• pp.1147-1165
• /
• 2010

In this paper we compare a torsion free sheaf F on $P^N$ and the free vector bundle $\oplus^n_{i=1}O_{P^N}(b_i)$ having same rank and splitting type. We show that the first one has always "less" global sections, while it has a higher second Chern class. In both cases bounds for the difference are found in terms of the maximal free subsheaves of F. As a consequence we obtain a direct, easy and more general proof of the "Horrocks' splitting criterion", also holding for torsion free sheaves, and lower bounds for the Chern classes $c_i$(F(t)) of twists of F, only depending on some numerical invariants of F. Especially, we prove for rank n torsion free sheaves on $P^N$, whose splitting type has no gap (i.e., $b_i{\geq}b_{i+1}{\geq}b_i-1$ 1 for every i = 1,$\ldots$,n-1), the following formula for the discriminant: $$\Delta(F):=2_{nc_2}-(n-1)c^2_1\geq-\frac{1}{12}n^2(n^2-1)$$. Finally in the case of rank n reflexive sheaves we obtain polynomial upper bounds for the absolute value of the higher Chern classes $c_3$(F(t)),$\ldots$,$c_n$(F(t)) for the dimension of the cohomology modules $H^iF(t)$ and for the Castelnuovo-Mumford regularity of F; these polynomial bounds only depend only on $c_1(F)$, $c_2(F)$, the splitting type of F and t.

• East Asian mathematical journal
• /
• v.9 no.2
• /
• pp.295-311
• /
• 1993

• Journal of the Chungcheong Mathematical Society
• /
• v.24 no.2
• /
• pp.319-335
• /
• 2011

We classify equivariant topoligical complex vector bundles over real projective plane under a compact Lie group (not necessarily effective) action. It is shown that nonequivariant Chern classes and isotropy representations at (at most) three points are sufficient to classify equivariant vector bundles over real projective plane except one case. To do it, we relate the problem to classification on two-sphere through the covering map because equivariant vector bundles over two-sphere have been already classified.

• Journal of the Korean Mathematical Society
• /
• v.52 no.3
• /
• pp.587-601
• /
• 2015

In this paper, we investigate the $\mathbb{R}$-complex Hermitian Finsler spaces, emphasizing the differences that separate them from the complex Finsler spaces. The tools used in this study are the Chern-Finsler and Berwald connections. By means of these connections, some classes of the $\mathbb{R}$-complex Hermitian Finsler spaces are defined, (e.g. weakly K$\ddot{a}$hler, K$\ddot{a}$hler, strongly K$\ddot{a}$hler). Here the notions of K$\ddot{a}$hler and strongly K$\ddot{a}$hler do not coincide, unlike the complex Finsler case. Also, some kinds of Berwald notions for such spaces are introduced. A special approach is devoted to obtain the equivalence conditions for an $\mathbb{R}$-complex Hermitian Finsler space to become a weakly Berwald or Berwald. Finally, we obtain the conditions under which an $\mathbb{R}$-complex Hermitian Finsler space with Randers metric is Berwald. We get some clear examples which illustrate the interest for this work.

• Journal of the Korean Mathematical Society
• /
• v.44 no.1
• /
• pp.35-54
• /
• 2007

Let $M_c$ = M(2, 0, c) be the moduli space of O(l)-semistable rank 2 torsion-free sheaves with Chern classes $c_1=0\;and\;c_2=c$ on a K3 surface X, where O(1) is a generic ample line bundle on X. When $c=2n\geq4$ is even, $M_c$ is a singular projective variety equipped with a holomorphic symplectic structure on the smooth locus. In particular, $M_c$ has trivial canonical divisor. In [22], O'Grady asks if there is any symplectic desingularization of $M_{2n}$ for $n\geq3$. In this paper, we show that there is no crepant resolution of $M_{2n}$ for $n\geq3$. This obviously implies that there is no symplectic desingularization.

• Journal of the Korean Mathematical Society
• /
• v.46 no.3
• /
• pp.589-607
• /
• 2009

In this paper we study arithmetically Cohen-Macaulay (ACM for short) vector bundles $\varepsilon$ of rank k $\geq$ 3 on hypersurfaces $X_r\;{\subset}\;{\mathbb{P}}^4$ of degree r $\geq$ 1. We consider here mainly the case of degree r = 4, which is the first unknown case in literature. Under some natural conditions for the bundle $\varepsilon$ we derive a list of possible Chern classes ($c_1$, $c_2$, $c_3$) which may arise in the cases of rank k = 3 and k = 4, when r = 4 and we give several examples.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.591-593
• /
• 2003

The study demonstrates the data extraction capabilities from the simulated ROCSAT-2 image by manual delineation. The GIS data are able to identify as fifteen categories of the classes I & II for the simulated ROCSAT-2 and SPOT image data. The areas of the paddy are identified almost the same results for both cases, but the arid farmland are identified differently about 30%. The ROCSAT-2 case can be also identified as seven more categories of the class III, but SPOT-4 case cannot.

• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.808-813
• /
• 2022

This study presents the problems of data imbalance, varying difficulties across target objects, and small objects in construction object segmentation for far-field monitoring and utilize compound loss functions to address it. Construction site scenes of assembling scaffolds were analyzed to test the effectiveness of compound loss functions for five construction object classes---workers, hardhats, harnesses, straps, hooks. The challenging problem was mitigated by employing a focal and Jaccard loss terms in the original loss function of LinkNet segmentation model. The findings indicates the importance of the loss function design for model performance on construction site scenes for far-field monitoring.