• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > Çѱ¹Á¤º¸°úÇÐȸ ³í¹®Áö > Á¤º¸°úÇÐȸ ³í¹®Áö B : ¼ÒÇÁÆ®¿þ¾î ¹× ÀÀ¿ë

Á¤º¸°úÇÐȸ ³í¹®Áö B : ¼ÒÇÁÆ®¿þ¾î ¹× ÀÀ¿ë

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) À§»ó°ü°è¸¦ À¯ÁöÇÏ´Â ¿ÏÈ­¹ýÀ» ÅëÇÑ Àü±â¿µµ¿ ¿µ»óÀÇ ¹ÝÁ¡ Á¤ÇÕ
¿µ¹®Á¦¸ñ(English Title) Matching Spots in Electrophoresis Images by Topology Preserving Relaxation
ÀúÀÚ(Author) Ȳ¿µ¼·   ÀÌÁ¾ÇÏ   Young-Sup Hwang   Jong-Ha Lee  
¿ø¹®¼ö·Ïó(Citation) VOL 39 NO. 06 PP. 0436 ~ 0443 (2012. 06)
Çѱ۳»¿ë
(Korean Abstract)
ÀÌÂ÷¿ø Àü±â¿µµ¿(2-DE) ¿µ»óÀÇ ¹ÝÁ¡Àº ´Ü¹éÁúÀ» ÀǹÌÇÑ´Ù. µÎ 2-DE ¿µ»óÀÇ ¹ÝÁ¡À» Á¤ÇÕÇÏ¸é ´Ü¹éÁúÀÇ »ý¼º°ú, ¼Ò¸ê, º¯ÀÌ µîÀ» ¾Ë ¼ö ÀÖ´Ù. Àü±â¿µµ¿ °úÁ¤¿¡¼­ Á¶Á÷ÀÇ »óÅÂ, ½ÇÇè Á¶°Ç¿¡ µû¶ó °°Àº ´Ü¹éÁúÀÌ¶óµµ 2-DE ¿µ»ó¿¡¼­ À§Ä¡°¡ Á¶±Ý¾¿ ´Ù¸£¹Ç·Î ¹ÝÁ¡ Á¤ÇÕ°ü°è´Â ±ÕÀÏÇÏÁö ¾ÊÀº ºñ¼±Çü º¯È¯ °ü°èÀÌ´Ù. ÅäÆú·ÎÁö¸¦ °í·ÁÇÏ¿© Á¤Á¡À» Á¤ÇÕÇÏ´Â ¹æ¹ýÀ¸·Î TPRL(Topology Preserving Relaxation Labeling) ¾Ë°í¸®ÁòÀ» °³¹ßÇÏ¿´´Ù. 2-DE ¿µ»óÀº ±¹ÁöÀûÀ¸·Î ÀÌ¿ôÇÑ ¹ÝÁ¡³¢¸®´Â ÅäÆú·ÎÁö°¡ À¯ÁöµÇ¹Ç·Î À̸¦ ¹Ý¿µÇÏ¿© TPRL ¾Ë°í¸®ÁòÀ» °³¼±ÇÏ¿© ´Ü¹éÁú ¹ÝÁ¡À» Á¤ÇÕÇÏ´Â ¹æ¾ÈÀ» Á¦½ÃÇÑ´Ù. °³¼±ÇÑ TPRL ¾Ë°í¸®ÁòÀº °¢Á¤Á¡¿¡¼­ shape context¸¦ °è»êÇÏ¿© Ãʱâ Á¤ÇÕ ºñ¿ëÀ» °è»êÇÏ°í, °Å¸®, °¢µµ¿Í ÆòźÇÔÀ¸·Î ÀûÇÕÁ¤µµ¸¦ °è»êÇÏ´Â °úÁ¤À¸·Î relaxationÀ» ¼öÇàÇÏ°í, Á¤ÇÕµµ¸¦ ³ôÀ̱â À§ÇÑ thin plate spline º¯È¯À» ¼öÇàÇÑ´Ù. ¼ö·ÅÇÒ ¶§±îÁö ÀÌ °úÁ¤À» ¹Ýº¹ÇÑ´Ù. ÇÕ¼º µ¥ÀÌÅÍ¿Í ½ÇÁ¦ µ¥ÀÌÅÍ¿¡ ´ëÇÏ¿© ½ÇÇèÀ» ÇÏ¿´°í, ½ÇÇè°á°ú ÀÌÀü¹æ¹ýº¸´Ù °³¼±µÊÀ» È®ÀÎÇÏ¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
A spot in a two dimensional electrophoresis (2-DE) image corresponds to a protein. Matching spots between two 2-DE images can make it possible to find out the generation, extinction and change of proteins. The location of the same protein can be different a little according to the tissue status and the conditions of experiment in the 2-DE process. The relation of spot matching is non-linear and non-uniform. We have developed the topology preserving relaxation labeling (TPRL) algorithm which can match points with their topology. Since the local topology of neighborhood spots preserves between 2-DE images, we improved the TPRL algorithm to match protein points by reflecting this characteristic. The improved TPRL algorithm computes the initial matching costs by the shape context at each spots, performs relaxation by computing the compatibility coefficient which represent the similarity between distance, angle and smoothness, and transforms the image by the thin plate spline model. This process iterates until convergence. We experimented the proposed algorithm on synthetic data and real data. The experimental results shows the improved matching of the proposed method.
Å°¿öµå(Keyword) ÀÌÂ÷¿ø Àü±â¿µµ¿(2-DE)   ÅäÆú·ÎÁö   ¹ÝÁ¡ Á¤ÇÕ   ¿ÏÈ­(relaxation)   ´Ü¹éÁú   2-Dimensional Electrophoresis   Topology   Point Matching   Relaxation   TPRL  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå