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

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > Çѱ¹ÀÎÅͳÝÁ¤º¸ÇÐȸ ³í¹®Áö

Çѱ¹ÀÎÅͳÝÁ¤º¸ÇÐȸ ³í¹®Áö

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) »ý¼ºÀû ´ë¸³½Ö ½Å°æ¸ÁÀ» ÀÌ¿ëÇÑ ±íÀÌÁöµµ ±â¹Ý ¿¬¹«Á¦°Å
¿µ¹®Á¦¸ñ(English Title) Single Image Dehazing Based on Depth Map Estimation via Generative Adversarial Networks
ÀúÀÚ(Author) ½Å°ÇÀ±   ±èµ¿¿í   È«¼º»ï   ÇÑ¸í¹¬   Gun-Yoon Shin   Dong-Wook Kim   Sung-sam Hong   Myung-Mook Han   ¿Õ¾ß¿À   Á¤¿ìÁø   ¹®¿µ½Ä   Yao Wang   Woojin Jeong   Young Shik Moon  
¿ø¹®¼ö·Ïó(Citation) VOL 19 NO. 05 PP. 0043 ~ 0054 (2018. 10)
Çѱ۳»¿ë
(Korean Abstract)
¿¬¹«°¡ ÀÖ´Â »óȲ¿¡¼­ ÃÔ¿µµÈ ¿µ»óÀº ³·Àº ´ëºñ·Î ÀÎÇØ ½ÃÀμºÀÌ ³·¾ÆÁö´Â ¹®Á¦°¡ ÀÖ´Ù. ÀÌ·¸°Ô ¿¬¹«·Î ÀÎÇØ È帴ÇÑ ¿µ»ó¿¡¼­ ¿¬¹«ÀÇ È¿°ú¸¦ Á¦°ÅÇÏ´Â °úÁ¤À» ¿¬¹«Á¦°Å¶ó°í ÇÑ´Ù. ¿¬¹«Á¦°Å¿¡¼­ °¡Àå Áß¿äÇÑ ¹®Á¦ Áß Çϳª´Â Àü´ÞÁöµµ (transmission map) ¶Ç´Â ±íÀÌÁöµµ (depth map)¸¦ Á¤È®ÇÏ°Ô ÃßÁ¤ÇÏ´Â °ÍÀÌ´Ù. º» ³í¹®¿¡¼­´Â Á¤È®ÇÑ ±íÀÌÁöµµ ÃßÁ¤À» À§ÇØ »ý¼ºÀû ´ë¸³½Ö ½Å°æ¸Á (Generative Adversarial Network: GAN)À» ÀÌ¿ëÇÑ Á¤È®ÇÑ ±íÀÌ ¿µ»ó ÃßÁ¤ ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. Á¦¾ÈµÈ GAN ¸ðµ¨Àº È帴ÇÑ ÀԷ¿µ»ó°ú ÀÌ¿¡ »óÀÀÇÏ´Â ±íÀÌÁöµµ °£ÀÇ ºñ¼±Çü ¸ÅÇÎÀ» ÇнÀÇÑ´Ù. ±×¸®°í ¿¬¹«Á¦°Å´Ü°è¿¡¼­´Â ÈÆ·ÃµÈ ¸ðµ¨À» »ç¿ëÇÏ¿© ÀԷ¿µ»óÀÇ ±íÀÌÁöµµ¸¦ ÃßÁ¤ÇÏ°í ÀÌ°ÍÀ» Àü´ÞÁöµµ¸¦ °è»êÇϴµ¥ »ç¿ëÇÑ´Ù. À̾ guided filter¸¦ »ç¿ëÇÏ¿© Àü´ÞÁöµµ¸¦ ´Ùµë´Â´Ù. ¸¶Áö¸·À¸·Î ´ë±â »ê¶õ ¸ðµ¨À» ±â¹ÝÀ¸·Î ¿¬¹«°¡ Á¦°ÅµÈ ¿µ»óÀ» º¹¿øÇÑ´Ù. Á¦¾ÈµÈ GAN ¸ðµ¨Àº ÇÕ¼º½Ç³»¿µ»óÀ¸·Î ÈƷõǾú´Ù. ÇÏÁö¸¸ ½ÇÁ¦ ¿¬¹«¿µ»ó¿¡ ´ëÇؼ­µµ Àû¿ëÇÒ ¼ö ÀÖ´Ù. À̸¦ ½ÇÇèÀ» ÅëÇØ Áõ¸íÇÏ¿´´Ù. ¶ÇÇÑ ½ÇÇè¿¡¼­ Á¦¾ÈµÈ ¹æ¹ýÀÌ ÀÌÀü¿¡ ¿¬±¸µÈ ¹æ¹ý¿¡ ºñÇØ ½Ã°¢Àû ¹× Á¤·®Àû Ãø¸é¿¡¼­ ¿ì¼öÇÑ °á°ú¸¦ ³ªÅ¸³Â´Ù.
¿µ¹®³»¿ë
(English Abstract)
Images taken in haze weather are characteristic of low contrast and poor visibility. The process of reconstructing clear-weather image from a hazy image is called dehazing. The main challenge of image dehazing is to estimate the transmission map or depth map for an input hazy image. In this paper, we propose a single image dehazing method by utilizing the Generative Adversarial Network(GAN) for accurate depth map estimation. The proposed GAN model is trained to learn a nonlinear mapping between the input hazy image and corresponding depth map. With the trained model, first the depth map of the input hazy image is estimated and used to compute the transmission map. Then a guided filter is utilized to preserve the important edge information of the hazy image, thus obtaining a refined transmission map. Finally, the haze-free image is recovered via atmospheric scattering model. Although the proposed GAN model is trained on synthetic indoor images, it can be applied to real hazy images. The experimental results demonstrate that the proposed method achieves superior dehazing results against the state-of-the-art algorithms on both the real hazy images and the synthetic hazy images, in terms of quantitative performance and visual performance.
Å°¿öµå(Keyword) ÀÛ¼ºÀÚ ½Äº°   ÀÛ¼ºÀÚ ºÐ¼®   ÇÕ¼º°ö ½Å°æ¸Á   ±â°èÇнÀ   ÄÚµå ºÐ¼®   Author Identification   Authorship Analysis   Convolutional Neural Network   Machine Learning   Code Analysis   ¿¬¹«Á¦°Å   ±íÀÌÁöµµ ÃßÁ¤   »ý¼ºÀû ´ë¸³½Ö ½Å°æ¸Á   dehaze   depth estimation   generative adversarial networks  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå