2019³â Ãß°èÇмú´ëȸ
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
µö·¯´×À» ÀÌ¿ëÇÑ ±âÈä, Æó·ÅÀÇ °ËÃâ°ú ºÐ·ù |
¿µ¹®Á¦¸ñ(English Title) |
Detection and classification of pneumothorax and pneumonia using deep learning |
ÀúÀÚ(Author) |
ÀÌÁ¾±Ù
±è¼±Áø
¾ÈÀçÇü
Jong-keun Lee
Seon-jin Kim
Jae-hyung Ahn
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 23 NO. 02 PP. 0638 ~ 0640 (2019. 10) |
Çѱ۳»¿ë (Korean Abstract) |
º» ³í¹®¿¡¼´Â ÈäºÎ x-¼±À» ÅëÇØ ±âÈä, Æó·ÅÀÇ Áø´Ü¿¡ µµ¿òÀ» ÁÙ ¼ö ÀÖ´Â µö·¯´× ¾Ë°í¸®ÁòÀ» Á¦¾ÈÇÑ´Ù. µö·¯´× ¾Ë°í¸®ÁòÀÇ ÇнÀ°ú Æò°¡¿¡´Â NIH(National Institute of Health) ÈäºÎ x-¼± »çÁø¿¡¼ ±âÈä, Æó·Å¿¡ ÇØ´çÇÏ´Â ¿µ»óÀ» »ç¿ëÇß´Ù. °ËÃâ°ú ºÐ·ù ¼º´ÉÀº AUROC(the Area Under the Receiver Operating Characteristic Curve)·Î Æò°¡Çß´Ù. Á¦¾ÈµÈ ¹æ¹ýÀº ±âÁ¸ÀÇ ÈäºÎ x-¼± °ËÃâ ¾Ë°í¸®Áòµé°ú ºñ±³ÇßÀ» ¶§ ±âÈäÀÇ °æ¿ì 0.02, Æó·ÅÀÇ °æ¿ì 0.17 ÀÌ»óÀÇ °³¼±µÈ ¼º´ÉÀ» º¸¿©ÁØ´Ù. |
¿µ¹®³»¿ë (English Abstract) |
In this paper, we propose the deep learning algorithm that can helps diagnose pneumothorax and pneumonia through chest x-ray images. The dataset consists of extracted pneumothorax and pneumonia from NIH(National Institute of Health) chest x-ray datum. The performance of the network was evaluated by comparing the AUROC(the Area Under the Receiver Operating Characteristic Curve). The proposed method is compared with conventional chest x-ray detection algorithms. and the results show improved performance of 0.02 in pneumothorax and 0.17 in pneumonia. |
Å°¿öµå(Keyword) |
µö·¯´×
¸Ó½Å·¯´×
±âÈä
Æó·Å
ÈäºÎ x-ray
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|