KCC 2021
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
µö¸Þµå³Ý: µö ·¯´× ±â¹Ý ÀÇ·á À̹ÌÁö ºÐÇÒ ¸ðµ¨ |
¿µ¹®Á¦¸ñ(English Title) |
DeepMedNet: Deep Learning based Medical Image Segmentation Model |
ÀúÀÚ(Author) |
Bekhzod Olimov
Barathi Subramanian
Jeonghong Kim
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 48 NO. 01 PP. 0576 ~ 0578 (2021. 06) |
Çѱ۳»¿ë (Korean Abstract) |
|
¿µ¹®³»¿ë (English Abstract) |
Deep convolutional neural networks (DCNN) assisted DL-based segmentation models to obtain state-of-the-art performance in the fields that are critical to human being, such as medicine. In this study, we propose a DCNN model (DeepMedNet) based on an elaborate pre-processing step and resourceful model architecture that obtains superior performance in both computational time and accuracy in comparison with the existing methods. Specifically, conducted experiments using the proposed method on two publicly available medical image datasets showed nearly 4x speed-up in training and inference time in comparison to the existing computationally expensive models and approximately 3% increase in image segmentation evaluation metrics compared with the currently available DL-based methods. |
Å°¿öµå(Keyword) |
computational efficiency
convolutional neural networks
medical image segmentation
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|