• 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¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)

Á¤º¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)

Current Result Document : 575 / 575

ÇѱÛÁ¦¸ñ(Korean Title) 1Â÷¿ø CNN¸¦ ÀÌ¿ëÇÑ ºñµð¿À Çൿ ÀÎ½Ä ¸ðµ¨ º´·Ä¼º Çâ»ó
¿µ¹®Á¦¸ñ(English Title) Improving Parallelism for Video Action Recognition Model Using One-dimensional Convolutional Neural Network
ÀúÀÚ(Author) ¼º¼öÁø   Â÷Á¤¿ø   Su-Jin Seong   Jeong-Won Cha  
¿ø¹®¼ö·Ïó(Citation) VOL 27 NO. 04 PP. 0216 ~ 0220 (2021. 04)
Çѱ۳»¿ë
(Korean Abstract)
µö·¯´× ÇÁ·¹ÀÓ¿öÅ©´Â ÄÄÇ»ÅÍ ºñÀü ¸¹Àº ºÐ¾ß¿¡¼­ °ý¸ñÇÒ ¸¸ÇÑ ¼º°ú¸¦ º¸¿©ÁÖ°í ÀÖ´Ù. ºñµð¿À Çൿ ÀÎ½Ä ºÐ¾ß ¿ª½Ã µö·¯´× ¸ðµ¨À» Àû¿ëÇϱâ À§ÇÑ ¸¹Àº ¿¬±¸µéÀÌ ¼öÇàµÇ¾ú´Ù. ÇÑ ¼±Ç࿬±¸´Â 2Â÷¿ø CNN À» ÀÌ¿ëÇØ °ø°£Àû ÇÇÃĸ¦ ÇнÀÇÏ°í À̸¦ RNN¿¡ ÀÔ·ÂÀ¸·Î Àü´ÞÇØ ÀÌ¿ëÇØ °ø°£Àû ÇÇÃÄ »çÀÌÀÇ ½Ã°£Àû »óÈ£ °ü°è¸¦ ÇнÀÇÏ´Â ¸ðµ¨ ±¸Á¶¸¦ Á¦¾ÈÇß´Ù. º» ³í¹®¿¡¼­´Â RNN ´ë½Å 1Â÷¿ø CNNÀ» ÀÌ¿ëÇØ ½Ã°£Àû »óÈ£ °ü°è¸¦ ÇнÀÇϵµ·Ï ¼±Çà ¿¬±¸ÀÇ ¸ðµ¨ ±¸Á¶¸¦ °³¼±ÇÏ´Â ¿¬±¸¸¦ ¼öÇàÇÑ´Ù. ÀÌ·¯ÇÑ ±¸Á¶ º¯°æÀ» ÅëÇØ RNNÀÇ ¼øÂ÷Àû ¿¬»ê °úÁ¤À» Á¦°ÅÇØ Çâ»óµÈ GPU È°¿ëµµ¸¦ ±â´ëÇÒ ¼ö ÀÖ´Ù. º» ³í¹®Àº ¼öÁ¤µÈ ¸ðµ¨ÀÌ Á¤È®µµ¸¦ ºñ½ÁÇÏ°Ô À¯ÁöÇϸ鼭 ¿¬»ê ½Ã°£ÀÌ ÁÙ¾îµå´Â °ÍÀ» º¸¿©ÁÖ´Â ½ÇÇè °á°ú¸¦ Á¦½ÃÇÔÀ¸·Î½á ÀÌ·¯ÇÑ ÁÖÀåÀ» µÞ¹Þħ ÇÑ´Ù.
¿µ¹®³»¿ë
(English Abstract)
The deep learning framework has shown remarkable results on numerous computer vision tasks. Many studies have been performed for video action recognition tasks to apply deep learning models to the task. One of the previous works suggested the model architecture, where spatial features are learned from 2D Convolutional Neural Networks (CNNs) and then passed to Recurrent Neural Networks (RNNs) to learn about temporal dependency among them. In this paper, we study the improved model architecture where the temporal relationship of spatial features is processed with 1D CNN instead of RNN. From this modification, we can expect better utilization of GPU by removing sequential operations of RNN. We support the argument based on the experiment results that show that it leads to the reduction in computation time and maintains a similar classification accuracy.
Å°¿öµå(Keyword) Æ®·£½ºÆ÷¸Ó ÀÎÄÚ´õ-µðÄÚ´õ ¸ðµ¨   ÀÚµ¿ Á¦¸ñ »ý¼º   ´Ü¾î ¼Õ½ÇÇÔ¼ö   ¹Ýº¹ Æä³ÎƼ   transformer encoder-decoder   automatic title generation   word loss   repeat penalty   ºñµð¿À ºÐ·ù   ºñµð¿À Çൿ ÀνĠ  1Â÷¿ø CNN   µö·¯´×   video classification   video action recognition   1D convolutional neural network   deep learning  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå