Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
º¹ºÎ CT ¿µ»ó¿¡¼ ÃéÀåÀÇ ºÒÈ®½Ç¼ºÀ» °í·ÁÇÑ °èÃþÀû ³×Æ®¿öÅ© ±â¹Ý ÀÚµ¿ ÃéÀå ºÐÇÒ |
¿µ¹®Á¦¸ñ(English Title) |
Automatic Pancreas Segmentation Based on Cascaded Network Considering Pancreatic Uncertainty in Abdominal CT Images |
ÀúÀÚ(Author) |
±èº¸Àº
À念Áø
±èÇмö
Boeun Kim
Youngjin Jang
Harksoo Kim
À±Çö´ã
±èÇöÁø
È«Çï·»
Hyeon Dham Yoon
Hyeonjin Kim
Helen Hong
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 48 NO. 05 PP. 0548 ~ 0555 (2021. 05) |
Çѱ۳»¿ë (Korean Abstract) |
ÃéÀå¾Ï °ËÃâ¿¡¼ÀÇ ÃéÀå ÇüÅ ÆľÇÀ» À§ÇØ º¹ºÎ CT ¿µ»ó¿¡¼ ÃéÀåÀ» ºÐÇÒÇÏ´Â °ÍÀÌ ÇÊ¿äÇÏ´Ù. º» ³í¹®¿¡¼´Â º¹ºÎ CT ¿µ»ó¿¡¼ ÃéÀåÀÇ À§Ä¡Àû, ÇüÅÂÀû ´Ù¾ç¼ºÀ¸·Î ÀÎÇØ ¹ß»ýÇÏ´Â ºÒÈ®½ÇÇÑ ¿µ¿ª¿¡ ´ëÇÑ Á¤º¸¸¦ ÇÔ²² °í·ÁÇÏ´Â DCNN ±â¹Ý ÃéÀå ÀÚµ¿ ºÐÇÒ ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. ù°, ¿µ»ó °£ ¹à±â°ª ¹× È¼Ò °ø°£ÀÇ Â÷À̸¦ ÁÙÀ̱â À§ÇØ ¹à±â°ª ¹× °ø°£ Á¤±Ôȸ¦ ¼öÇàÇÑ´Ù. µÑ°, »ï´Ü¸é U-Net ±â¹Ý 2.5Â÷¿ø ºÐÇÒ ³×Æ®¿öÅ© ¹× ´Ù¼ö ÅõÇ¥¸¦ ÅëÇØ ÃéÀåÀ» À§Ä¡ÈÇÑ´Ù. ¼Â°, À§Ä¡ÈµÈ 3Â÷¿ø ÃéÀå ¿µ¿ª¿¡¼ ÃéÀåÀÇ ºÒÈ®½ÇÇÑ ¿µ¿ª Á¤º¸¸¦ °í·ÁÇÏ´Â U-Net ±â¹Ý 3Â÷¿ø ºÐÇÒ ³×Æ®¿öÅ©¸¦ ÀÌ¿ëÇÏ¿© ÃéÀåÀ» ºÐÇÒÇÑ´Ù. Á¦¾È¹æ¹ýÀ» ÅëÇÑ ºÐÇÒ °á°úÀÇ DSC´Â 83.50%·Î, Ⱦ´Ü¸é, °ü»ó¸é, ½Ã»ó¸é¿¡¼ 2Â÷¿ø U-NetÀ» ÀÌ¿ëÇÑ ºÐÇÒ, 2.5Â÷¿ø ºÐÇÒ ¹× À§Ä¡ÈµÈ ¿µ¿ª¿¡¼ 3Â÷¿ø U-NetÀ» ÀÌ¿ëÇÑ ºÐÇÒ ¹æ¹ý ´ëºñ °¢°¢ 10.30%p, 10.44%p, 6.52%p, 1.14%p, 3.95%p Çâ»óµÇ¾ú´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
Pancreas segmentation from abdominal CT images is a prerequisite step for understanding the shape of the pancreas in pancreatic cancer detection. In this paper, we propose an automatic pancreas segmentation method based on a deep convolutional neural network(DCNN) that considers information about the uncertain regions generated by the positional and morphological diversity of the pancreas in abdominal CT images. First, intensity and spacing normalizations are performed in the whole abdominal CT images. Second, the pancreas is localized using 2.5D segmentation networks based on U-Net on the axial, coronal, and sagittal planes and by combining through a majority voting. Third, pancreas segmentation is performed in the localized volume using a 3D U-Net-based segmentation network that takes into account the information about the uncertain areas of the pancreas. The average DSC of pancreas segmentation was 83.50%, which was 10.30%p, 10.44%p, 6.52%p, 1.14%p, and 3.95%p higher than the segmentation method using 2D U-Net at axial view, coronal view, sagittal view, majority voting of the three planes, and 3D U-Net at localized volume, respectively.
|
Å°¿öµå(Keyword) |
Á¤´ä È帱º ŽÁö
Áú¹® »ý¼º
ÁúÀÇÀÀ´ä
µ¥ÀÌÅÍ ÀÚµ¿ ±¸Ãà
Span Matrix
2´Ü°è ÇнÀ
answer candidates detection
question generation
question-answering
automatic data construction
span matrix
2-step learning
µö ÄÁº¼·ç¼Ç ½Å°æ¸Á(DCNN)
ÃéÀå ºÐÇÒ
º¹ºÎ CT ¿µ»ó
U-Net
ºÒÈ®½Ç¼º
deep convolutional neural network (DCNN)
pancreas segmentation
abdominal CT images
U-Net
uncertainty
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|