• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > Çѱ¹Á¤º¸°úÇÐȸ ³í¹®Áö > Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)

Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) ¹è¼ö ÆØâµÈ ÄÁº¼·ç¼Ç°ú Ư¡Çհ踦 ÀÌ¿ëÇÑ °´Ã¼ °ËÃ⠹麻 ³×Æ®¿öÅ©
¿µ¹®Á¦¸ñ(English Title) Backbone Network for Object Detection with Multiple Dilated Convolutions and Feature Summation
ÀúÀÚ(Author) ¹Ù´Ï ³ªÅ»¸®¾Æ ÄïÆ®Á¶³ë   °í½ÂÇö   ¹æ¾ç   Á¶±Ù½Ä   Vani Natalia Kuntjono   Seunghyun Ko   Yang Fang   Geunsik Jo  
¿ø¹®¼ö·Ïó(Citation) VOL 45 NO. 08 PP. 0786 ~ 0791 (2018. 08)
Çѱ۳»¿ë
(Korean Abstract)
ÄÁº¼·ç¼Ç ´º·² ³×Æ®¿öÅ©ÀÇ ¹ßÀüÀ¸·Î ÀÎÇØ °´Ã¼ ŽÁö, À̹ÌÁö ¼¼ºÐÈ­ ¹× °´Ã¼ ºÐ·ù ºÐ¾ß¿¡¼­µµ 100°³ ÀÌ»óÀÇ ÄÁº¼·ç¼Ç ·¹À̾ »ç¿ëÇÏ´Â Deep CNNÀ» »ç¿ëÇÏ´Â Ãß¼¼·Î À̾îÁö°í ÀÖ´Ù. ±×·¯³ª Deep CNNÀ» »ç¿ëÇϱâ À§ÇØ ¸¹Àº ±×·¡ÇÈ ¸Þ¸ð¸®°¡ ÇÊ¿äÇϸç Á¦ÇÑµÈ ÀÚ¿øÀ̳ª ½Ç½Ã°£ °´Ã¼ ŽÁö¸¦ ¿øÇÏ´Â »ç¶÷µé¿¡°Ô´Â ÀÌ·± Deep CNNÀÌ ÀûÇÕÇÏÁö ¾Ê´Ù. º» ³í¹®¿¡¼­´Â ¹è¼ö ÆØâµÈ ÄÁº¼·ç¼Ç°ú Ư¡ÇÕ°è ±â¹ÝÀÇ °´Ã¼ ŽÁö¸¦ À§ÇÑ »õ·Î¿î ¹éº» ³×Æ®¿öÅ©¸¦ Á¦¾ÈÇÑ´Ù. Ư¡ÇÕ°è´Â ±×·¡µð¾ðÆ®¸¦ ½±°Ô Àü´ÞÇÏ°í ÄÁº¼ºùÀ¸·Î ÀÎÇØ ¹ß»ýÇÏ´Â °ø°£ Á¤º¸ÀÇ ¼Õ½ÇÀ» ÃÖ¼ÒÈ­ÇÑ´Ù. ±×¸®°í ÆØâµÈ ÄÁº¼·ç¼ÇÀ» »ç¿ëÇÔÀ¸·Î½á º¯¼ö¸¦ Ãß°¡ÇÏÁö ¾Ê°íµµ °³º° ´º·±ÀÇ ¼ö¿ë ¿µ¿ªÀ» ³ÐÈú ¼ö ÀÖ´Ù. ¶ÇÇÑ, DeepÇÏÁö ¾ÊÀº ´º·² ³×Æ®¿öÅ©¸¦ ¹éº»À¸·Î »ç¿ëÇÔÀ¸·Î½á Á¦ÇÑµÈ ÀÚ¿øÀ¸·Î À̹ÌÁö³Ý µ¥ÀÌÅÍ ¼¼Æ®¿¡¼­ »çÀü ±³À°À» ÇÏÁö ¾Ê´õ¶óµµ Á¦¾ÈÇÏ´Â ³×Æ®¿öÅ©¸¦ »ç¿ëÇÒ ¼ö ÀÖ´Ù. Pascal VOC ¹× MS COCO µ¥ÀÌÅ͸¦ »ç¿ëÇÑ ½ÇÇè °á°ú Á¦¾ÈµÈ ³×Æ®¿öÅ©´Â °¢°¢ 71%¿Í 38.2%ÀÇ Á¤È®µµ¸¦ º¸¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
The advancement of CNN leads to the trend of using very deep convolutional neural network which contains more than 100 layers not only for object detection, but also for image segmentation and object classification. However, deep CNN requires lots of resources, and so is not suitable for people who have limited resources or real time requirements. In this paper, we propose a new backbone network for object detection with multiple dilated convolutions and feature summation. Feature summation enables easier flow of gradients and minimizes loss of spatial information that is caused by convolving. By using multiple dilated convolution, we can widen the receptive field of individual neurons without adding more parameters. Furthermore, by using a shallow neural network as a backbone network, our network can be trained and used in an environment with limited resources and without pre-training it in ImageNet dataset. Experiments demonstrate we achieved 71% and 38.2% of accuracy on Pascal VOC and MS COCO dataset, respectively.
Å°¿öµå(Keyword) °´Ã¼ °ËÃâ   ¹éº» ³×Æ®¿öÅ©   ¹è¼ö ÆØâ µÈ ÄÁº¼·ç¼Ç   Ư¡ÇÕ°è   object detection   backbone network   multiple dilated convolutions   feature summation  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå