• 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¸Ã³¸®ÇÐȸ ³í¹®Áö > Á¤º¸Ã³¸®ÇÐȸ ³í¹®Áö ¼ÒÇÁÆ®¿þ¾î ¹× µ¥ÀÌÅÍ °øÇÐ

Á¤º¸Ã³¸®ÇÐȸ ³í¹®Áö ¼ÒÇÁÆ®¿þ¾î ¹× µ¥ÀÌÅÍ °øÇÐ

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) Àû´ëÀû »ý¼º ½Å°æ¸ÁÀ» ÅëÇÑ ¾ó±¼ ºñµð¿À ½ºÅ¸ÀÏ ÇÕ¼º ¿¬±¸
¿µ¹®Á¦¸ñ(English Title) Style Synthesis of Speech Videos Through Generative Adversarial Neural Networks
ÀúÀÚ(Author) ÃÖÈñÁ¶   ¹Ú±¸¸¸   Choi Hee Jo   Park Goo Man  
¿ø¹®¼ö·Ïó(Citation) VOL 11 NO. 11 PP. 0465 ~ 0472 (2022. 11)
Çѱ۳»¿ë
(Korean Abstract)
º» ¿¬±¸¿¡¼­´Â ±âÁ¸ÀÇ µ¿¿µ»ó ÇÕ¼º ³×Æ®¿öÅ©¿¡ ½ºÅ¸ÀÏ ÇÕ¼º ³×Æ®¿öÅ©¸¦ Á¢¸ñ½ÃÄÑ µ¿¿µ»ó¿¡ ´ëÇÑ ½ºÅ¸ÀÏ ÇÕ¼ºÀÇ ÇÑ°èÁ¡À» ±Øº¹ÇÏ°íÀÚ ÇÑ´Ù. º» ³í¹®ÀÇ ³×Æ®¿öÅ©¿¡¼­´Â µ¿¿µ»ó ÇÕ¼ºÀ» À§ÇØ ½ºÅ¸ÀÏ°µ ÇнÀÀ» ÅëÇÑ ½ºÅ¸ÀÏ ÇÕ¼º°ú µ¿¿µ»ó ÇÕ¼º ³×Æ®¿öÅ©¸¦ ÅëÇØ ½ºÅ¸ÀÏ ÇÕ¼ºµÈ ºñµð¿À¸¦ »ý¼ºÇϱâ À§ÇØ ³×Æ®¿öÅ©¸¦ ÇнÀ½ÃŲ´Ù. Àι°ÀÇ ½Ã¼±À̳ª Ç¥Á¤ µîÀÌ ¾ÈÁ¤ÀûÀ¸·Î ÀüÀ̵DZ⠾î·Á¿î Á¡À» °³¼±Çϱâ À§ÇØ 3Â÷¿ø ¾ó±¼ º¹¿ø±â¼úÀ» Àû¿ëÇÏ¿© 3Â÷¿ø ¾ó±¼ Á¤º¸¸¦ ÀÌ¿ëÇÏ¿© ¸Ó¸®ÀÇ Æ÷Áî¿Í ½Ã¼±, Ç¥Á¤ µîÀÇ Áß¿äÇÑ Æ¯Â¡À» Á¦¾îÇÑ´Ù. ´õºÒ¾î, ÇìµåÅõÇìµå++ ³×Æ®¿öÅ©ÀÇ ¿ªµ¿¼º, ÀÔ ¸ð¾ç, À̹ÌÁö, ½Ã¼± 󸮿¡ ´ëÇÑ ÆǺ°±â¸¦ °¢°¢ ÇнÀ½ÃÄÑ °³¿¬¼º°ú ÀÏ°ü¼ºÀÌ ´õ¿í À¯ÁöµÇ´Â ¾ÈÁ¤ÀûÀÎ ½ºÅ¸ÀÏ ÇÕ¼º ºñµð¿À¸¦ »ý¼ºÇÒ ¼ö ÀÖ´Ù. ÆäÀ̽º Æ÷·»½Ä µ¥ÀÌÅͼ°ú ¸ÞÆ®·ÎÆú¸®Åº ¾ó±¼ µ¥ÀÌÅͼÂÀ» ÀÌ¿ëÇÏ¿© ´ë»ó ¾ó±¼ÀÇ ÀÏ°üµÈ ¿òÁ÷ÀÓÀ» À¯ÁöÇϸ鼭 ´ë»ó ºñµð¿À·Î º¯È¯ÇÏ¿©, Àڱ⠾󱼿¡ ´ëÇÑ 3Â÷¿ø ¾ó±¼ Á¤º¸¸¦ ÀÌ¿ëÇÑ ºñµð¿À ÇÕ¼ºÀ» ÅëÇØ ÀÚ¿¬½º·¯¿î µ¥ÀÌÅ͸¦ »ý¼ºÇÏ¿© ¼º´ÉÀ» Áõ°¡½ÃÅ´À» È®ÀÎÇß´Ù.
¿µ¹®³»¿ë
(English Abstract)
In this paper, the style synthesis network is trained to generate style-synthesized video through the style synthesis through training Stylegan and the video synthesis network for video synthesis. In order to improve the point that the gaze or expression does not transfer stably, 3D face restoration technology is applied to control important features such as the pose, gaze, and expression of the head using 3D face information. In addition, by training the discriminators for the dynamics, mouth shape, image, and gaze of the Head2head network, it is possible to create a stable style synthesis video that maintains more probabilities and consistency. Using the FaceForensic dataset and the MetFace dataset, it was confirmed that the performance was increased by converting one video into another video while maintaining the consistent movement of the target face, and generating natural data through video synthesis using 3D face information from the source video¡¯s face.
Å°¿öµå(Keyword) Àû´ëÀû »ý¼º ³×Æ®¿öÅ©   ºñµð¿À »ý¼º   ½ºÅ¸ÀÏ º¯È¯   ½ºÅ¸ÀÏ ÇÕ¼º ³×Æ®¿öÅ©   µ¿¿µ»ó ÇÕ¼º ³×Æ®¿öÅ©   Generative Adversarial Network   Video Generation   Style Transfer   Style Synthesis Network   Video Synthesis Network  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå