Àá½Ã¸¸ ±â´Ù·Á ÁÖ¼¼¿ä. ·ÎµùÁßÀÔ´Ï´Ù.

Àü»ê󸮸¦ ÅëÇÑ LinacgramÀÇ È­Áú°³¼± Enhancement of Image Contrast in Linacgram through Image Processing

´ëÇѹæ»ç¼±Á¾¾çÇÐȸÁö 2000³â 18±Ç 4È£ p.345 ~ 354
¼Ò¼Ó »ó¼¼Á¤º¸
¼­Çö¼÷/Hyun Suk Suh ½ÅÇö±³/ÀÌ·¹³ª/Hyun Kyo Shin/Rena Lee

Abstract

¸ñÀû: ¹æ»ç¼±Á¶»ç¾ß¸¦ È®ÀÎÇÏ´Â º¸ÆíÀûÀÎ ¹æ¹ýÀÎ linacgramÀº Àú´ëÁ¶µµ(low contrast)ÀÇ ¿µ»óÀ» º¸¿©ÁÖ°í ÀÖ¾î Á¤È®ÇÑ ¿µ»óÀ» È®ÀÎÇϴµ¥ ¹®Á¦Á¡ÀÌ ÀÖ´Ù. µû¶ó¼­ º» ¿¬±¸´Â linacgramÀÇ ´ëÁ¶µµ¸¦ ³ôÀÌ´Â Àú°¡Çü È®Àιæ¹ýÀ» ¸ð»öÇÏ¿© ¿µ»óÆǵ¶°ú Á¶»ç¾ß È®Àο¡
µµ¿òÀÌ
µÇ°íÀÚ ÇÑ´Ù.
´ë»ó ¹× ¹æ¹ý: ÀÎü¸ðÇüÀ» »ç¿ëÇÏ¿© ¾ò¾îÁø Çʸ§ ¿µ»óÀ» Çʸ§Àü¿ë ½ºÄ³³Ê(Diagnostic Pro)¸¦
ÅëÇØ Optical Density Scan, Histogram Equalized, Linear Histogram Based (HB), Linear
Histogram Independent, Linear Optical density (OD), Logarithmic ¹× Power, Square Root
scan ¹æ½ÄÀ¸·Î µðÁöÅÐ È­ ÇÏ¿´´Ù. °¢±â ´Ù¸¥ ¹æ½ÄÀ¸·Î Àü»ê ÀÔ·ÂµÈ ¿µ»óÀÇ ½ÅÈ£ ºÐÆ÷µµ¸¦
¾ò¾î signal intensity¸¦ ºñ±³ÇÑ ÈÄ pallette fitting ¹æ½ÄÀ» ÅëÇØ ¿µ»óÀ» À籸¼ºÇÏ¿´°í À籸¼º
µÈ ¿µ»óÀ» ºñ±³ºÐ¼®ÇÏ¿´´Ù. ½ÇÁ¦ Ä¡·á¿¡¼­ ¾ò¾îÁø °¢ ÀÎü ºÎÀ§º° linacgramµµ µ¿ÀÏÇÑ ¹æ¹ý
À¸·Î ó¸®ÇÑ ÈÄ È­Áú °³¼±µµ¸¦ ¾Ë¾Æº¸¾Ò´Ù.
°á°ú: ÀÎü¸ðÇüÀ» ÅëÇØ ¾ò¾îÁø ¿µ»óÀÇ ½ÅÈ£ ºÐÆ÷¿µ¿ªÀº Logarithmic ¹æ½ÄÀ» ¼±ÅÃÇßÀ» ¶§ ÃÖ
¼Ò°ªÀÎ 3192°¡ ³ª¿Ô°í Square Root¹æ½ÄÀ» »ç¿ëÇßÀ» ¶§ ÃÖ´ë°ªÀÎ 21940°¡ ³ª¿Ô´Ù. ÀÌ·¯ÇÑ °ª
µéÀ» ¸ð´ÏÅÍ »ó¿¡¼­ ±¸ÇöÇÒ ¼ö ÀÖ´Â 256 gray scale·Î ¹Ù²Ù¾î º¸¾ÒÀ» ¶§ 7~30% ¸¸ »ç¿ëµÇ
¾îÁö°í ÀÖÀ½À» ¾Ë¼ö ÀÖ¾ú´Ù. Pallette fitting ¹æ½ÄÀ» ÅëÇÏ¿© ¸ð´ÏÅÍÀÇ ÃÖ´ëÇ¥Çö °ªÀÎ 256 °è
Á¶µµ·Î Gray Scale Expansion (GSE) ÇÔÀ¸·Î½á ¸ð´ÏÅÍ°¡ Áö¿øÇÏ´Â 8bit gray scale pallette
ÀÇ Àü ¹üÀ§¸¦ »ç¿ëÇÏ¿© ´ëÁ¶µµ°¡ °³¼±µÇ¾ú´Ù. ÀÓ»ó¿¡¼­ ¾ò¾îÁø °¢ ÀÎü ºÎÀ§º° ¹«¸­°üÀý, µÎ
°æºÎ, Æó, °ñ¹æ¿µ»ó¿¡¼­µµ GSE ó¸®ÇÏ¿© ¾ò¾îÁø ¿µ»óÀÌ ÇغÎÇÐÀû ±¸Á¶¸¦ Æǵ¶Çϴµ¥ µµ¿ò
ÀÌ µÇ¾ú´Ù.
°á·Ð: GSE ¿µ»óÀÇ À籸¼ºÀº ´ëÁ¶µµ¸¦ Áõ°¡ ½Ãų»Ó ¾Æ´Ï¶ó ÀÎü³» °ü½ÉºÎÀ§ÀÇ ³óµµºÐÆ÷¸¦
º°µµ·Î À籸¼ºÇÒ ¼ö ÀÖÀ¸¹Ç·Î ÀÌÁß¹æ»ç¼±Á¶»ç(double exposure)¿¡ ÀÇÇØ ¹ß»ýµÇ´Â È­ÁúÀÇ Àú
Çϸ¦ º¸Á¤ÇÔÀ¸·Î½á È­Áú °³¼±À» °¡´ÉÇÏ°Ô ÇÏ¿´´Ù. Linacgram È­Áú °³¼±Àº simulation image
¹× Ä¡·á°èȹ¿¡¼­ ¹ß»ýÇÑ DDR°ú multi-layer Áßø¿µ»ó ºÐ¼®¿¡ »ç¿ëÇÒ ¼ö ÀÖÀ¸¸ç ¿µ»ó ºñ±³
½Ã Ä¡·áºÎÀ§ÀÇ ½Å¼ÓÇÏ°í Á¤¹ÐÇÑ È®ÀÎÀ» °¡´ÉÇÏ°Ô ÇÏ¿´´Ù.

Purpose: Conventional radiation therapy portal images gives low contrast images. The purpose of this study was to enhance image contrast of a linacgram by developing a low-cost image processing method.
Materials and Methods: Chest linacgram was obtained by irradiating humanoid phantom
and scanned using Diagnostic-Pro scanner for image processing. Several types of scan
method were used in scanning. These include optical density scan, histogram equalized
scan, linear histogram based scan, linear histogram independent scan, linear optical
density scan, logarithmic scan, and power square root scan. The histogram distribution
of the scanned images were plotted and the ranged of the gray scale were compared
among various scan types. The scanned images were then transformed to the gray
window by pallette fitting method and the contrast of the reprocessed portal images
were evaluated for image improvement. Portal images of patients were also taken at
various anatomic sites and the images were processed by Gray Scale Expansion (GSE)
method. The patient images were analyzed to examine the feasibility of using the GSE
technique in clinic.
Results: The histogram distribution showed that minimum and maximum gray scale
ranges of 3192 and 21940 were obtained when the image was scanned using logarithmic
method and square root method, respectively. Out of 256 gray scale, only 7 to 30% of
the steps were used. After expanding the gray scale to full range, contrast of the portal
images were improved. Experiment performed with patient image showed that improved
identification of organs were achieved by GSE in portal images of knee joint, head and
neck, lung, and pelvis.
Conclusion: Phantom study demonstrated that the GSE technique improved image
contrast of a linacgram. This indicates that the decrease in image quality resulting from
the dual exposure, could be improved by expanding the gray scale. As a result, the
improved technique will make it possible to compare the digitally reconstructed
radiographs(DDR) and simulation image for evaluating the patient positioning error.

Å°¿öµå

Linacgram; Portal Imaging; ¿µ»óó¸®; ´ëÁ¶µµ; Linacgram; Portal image; Image processing; Constrast;

¿ø¹® ¹× ¸µÅ©¾Æ¿ô Á¤º¸

   

µîÀçÀú³Î Á¤º¸

KCI
KoreaMed
KAMS