Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
´Ù´Ü°è Seg-Unet ¸ðµ¨À» ÀÌ¿ëÇÑ ¹æ»ç¼± »çÁø¿¡¼ÀÇ End-to-end °ñ Á¾¾ç ºÐÇÒ ¹× ºÐ·ù |
¿µ¹®Á¦¸ñ(English Title) |
End-to-end Bone Tumor Segmentation and Classification from X-ray Images by Using Multi-level Seg-Unet Model |
ÀúÀÚ(Author) |
µµ´©µûÀÌ
Á¤¼ºÅÃ
¾çÇüÁ¤
±è¼öÇü
Nhu-Tai Do
Sung-Taek Jung
Hyung-Jeong Yang
Soo-Hyung Kim
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 47 NO. 02 PP. 0170 ~ 0179 (2020. 02) |
Çѱ۳»¿ë (Korean Abstract) |
¹«¸ °ñÁ¾¾ç °ËÃâÀº ÀÇ·áÁø´Ü º¸Á¶ ½Ã½ºÅÛ ±¸Çö¿¡ ÀÖ¾î¼ Áß¿äÇÑ ¿ªÇÒÀ» ´ã´çÇÑ´Ù. Áö±Ý±îÁö Á¦½ÃµÈ ¹æ¹ý Áß ÀÔ·Â X-ray ¿µ»ó¿¡¼ Á¾¾çÀ» °ËÃâÇÏ°í À̸¦ ºÐ·ùÇÏ´Â ±â´ÉÀÌ ¸ðµÎ Æ÷ÇÔµÈ end-to-end ½Ã½ºÅÛÀº ¾ø´Ù. º» ³í¹®¿¡¼´Â ´ÙÁß µö·¯´×¿¡ ±â¹ÝÇÑ end-to-end ½Ã½ºÅÛÀ» Á¦¾ÈÇÑ´Ù. À̸¦ À§ÇØ ¿ì¸®´Â ¿µ»ó³» Á¾¾çºÎºÐ¿¡ ´ëÇÑ °Å¸®º¯È¯À¸·ÎºÎÅÍ ´Ù´Ü°è ¸¶½ºÅ©¸¦ »ý¼ºÇÏ°í, À̸¦ ÇØ´ç Á¾¾çÀÇ Àǹ̷ÐÀû Á¤º¸¸¦ ÃßÃâÇÏ´Â ½Å°æ¸ÁÀÇ guided filter·Î È°¿ëÇÑ´Ù. ¶ÇÇÑ, Á¦¾ÈµÈ ½Å°æ¸Á ±¸Á¶´Â Á¾¾çÀÇ ºÐÇÒ°ú ºÐ·ù °úÁ¤À» ÇнÀÇÏ´Â °úÁ¤¿¡¼ Á¤±ÔÈÇÏ´Â È¿°ú¸¦ Æ÷ÇÔÇÏ°í ÀÖ´Ù. Á¦¾ÈµÈ ½Å°æ¸Á ¸ðµ¨ÀÌ Àü³²´ëÇб³º´¿ø¿¡¼ ±¸ÃàÇÑ µ¥ÀÌÅͼ¿¡ ´ëÇØ ´Ù¸¥ ±â¹ýµéº¸´Ù ¿ì¼öÇÑ ¼º´ÉÀ» º¸ÀÓÀ» ÀÔÁõÇÏ¿´´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
Knee bone tumor detection plays an essential role in assisting the clinical diagnosis process. To the best of our knowledge, there is no method to integrate end-to-end segmentation and classification for this problem. In this paper, we propose a multi-task deep learning architecture for classification and segmentation of the tumor regions in the knee bone. Also, we introduce multi-level distance masks from the distance transform of tumor region, and these multi-level distance masks have a role as a guided filter in enabling the network to capture semantic data around tumor regions. Besides, the architecture has a regularizing effect on the learning process between segmentation and classification. Our model was evaluated on the Chonnam National University Hospital dataset and achieved good performance compared to other methods.
|
Å°¿öµå(Keyword) |
µö·¯´×
¹«¸ °ñ Á¾¾ç
ÀÇ·á ¿µ»ó ºÐÇÒ
unet
segnet
dice loss
deep learning
knee bone tumor
medical image segmentation
unet
segnet
dice loss
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|