Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
½Ç½Ã°£ 3Â÷¿ø ÀÚ¼¼ ¹× ¿ÜÇü º¹¿øÀ» À§ÇÑ ¿ª±â±¸ÇÐ ±â¹ÝÀÇ ºÐ»êÇü ÇÁ·¹ÀÓ¿öÅ© |
¿µ¹®Á¦¸ñ(English Title) |
A Distributed Framework for Realtime 3D Human Pose and Shape Reconstruction Based on Inverse Kinematics |
ÀúÀÚ(Author) |
ȲŹÎ
±èÁöÀº
±è¹ÎÁØ
Tae-min Hwang
Ji-eun Kim
Min-joon Kim
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 45 NO. 01 PP. 2131 ~ 2133 (2022. 06) |
Çѱ۳»¿ë (Korean Abstract) |
|
¿µ¹®³»¿ë (English Abstract) |
As the growth of the market for 3D multimedia contents, several researches for 3D human pose estimation from multiple views has been studied. However, 3D keypoint estimation methods which relies on deep neural networks cannot reconstruct realistic body structure due to the lack of the body-part rotations. To address this problem, we propose a framework for 3D human pose and shape reconstruction based on inverse kinematics.
|
Å°¿öµå(Keyword) |
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|