Á¤º¸Ã³¸®ÇÐȸ ³í¹®Áö ¼ÒÇÁÆ®¿þ¾î ¹× µ¥ÀÌÅÍ °øÇÐ
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
3Â÷¿ø ÀÚ¼¼ ÃßÁ¤À» À§ÇÑ µö·¯´× ±â¹Ý ÀÌ»óÄ¡ °ËÃâ ¹× º¸Á¤ ±â¹ý |
¿µ¹®Á¦¸ñ(English Title) |
Deep Learning-Based Outlier Detection and Correction for 3D Pose Estimation |
ÀúÀÚ(Author) |
¹Ú¼ÒÇö
±è¼¿¬
¹Ú¿µÈ£
So-Hyun Park
Seo-Yeon Kim
Young-Ho Park
ÁÖÂù¾ç
¹ÚÁö¼º
À̵¿È£
Chan-Yang Ju
Ji-Sung Park
Dong-Ho Lee
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 11 NO. 10 PP. 0419 ~ 0426 (2022. 10) |
Çѱ۳»¿ë (Korean Abstract) |
º» ³í¹®¿¡¼´Â ´Ù¾çÇÑ ¿îµ¿ ¸ð¼Ç¿¡¼ 3Â÷¿ø »ç¶÷ ÀÚ¼¼ ÃßÁ¤ ¸ðµ¨ÀÇ Á¤È®µµ¸¦ Çâ»óÇÏ´Â ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. ±âÁ¸ÀÇ »ç¶÷ ÀÚ¼¼ ÃßÁ¤ ¸ðµ¨Àº »ç¶÷ÀÇ ÀÚ¼¼¸¦ ÃßÁ¤ÇÒ ¶§ ÁÂÇ¥ ¿ÀÂ÷¸¦ À¯¹ßÇÏ´Â Èçµé¸², ¹ÝÀü, ±³È¯, ¿À°ËÃâ µîÀÇ ¹®Á¦°¡ ¹ß»ýÇÑ´Ù. ÀÌ·¯ÇÑ ¹®Á¦´Â »ç¶÷ ÀÚ¼¼ ÃßÁ¤ ¸ðµ¨ÀÇ Á¤È®ÇÑ ÀÚ¼¼ ÃßÁ¤À» ¾î·Æ°Ô ÇÑ´Ù. À̸¦ ÇØ°áÇϱâ À§ÇØ º» ³í¹®¿¡¼´Â µö·¯´× ±â¹Ý ÀÌ»óÄ¡ °ËÃâ ¹× º¸Á¤ ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. µö·¯´× ±â¹ÝÀÇ ÀÌ»óÄ¡ °ËÃâ ¹æ¹ýÀº ¿©·¯ ¸ð¼Ç¿¡¼ ÁÂÇ¥ÀÇ ÀÌ»óÄ¡¸¦ È¿°úÀûÀ¸·Î °ËÃâÇÏ°í, ¸ð¼ÇÀÇ Æ¯Â¡À» È°¿ëÇÑ ±ÔÄ¢ ±â¹Ý º¸Á¤ ¹æ¹ýÀ» ÅëÇØ ÀÌ»óÄ¡¸¦ º¸Á¤ÇÑ´Ù. ´Ù¾çÇÑ ½ÇÇè°ú ºÐ¼®À» ÅëÇÏ¿© Á¦¾ÈÇÏ´Â ¹æ¹ýÀÌ °ñÇÁ ½ºÀ® ¸ð¼Ç°ú ´Ù¾çÇÑ ¿îµ¿ ¸ð¼Ç¿¡¼µµ »ç¶÷ÀÇ ÀÚ¼¼¸¦ Á¤È®È÷ ÃßÁ¤ÇÒ ¼ö ÀÖ°í, 3Â÷¿ø ÁÂÇ¥ µ¥ÀÌÅÍ¿¡¼µµ È®Àå °¡´ÉÇÔÀ» º¸ÀδÙ. |
¿µ¹®³»¿ë (English Abstract) |
In this paper, we propose a method to improve the accuracy of 3D human pose estimation model in various move motions. Existing human pose estimation models have some problems of jitter, inversion, swap, miss that cause miss coordinates when estimating human poses. These problems cause low accuracy of pose estimation models to detect exact coordinates of human poses. We propose a method that consists of detection and correction methods to handle with these problems. Deep learning-based outlier detection method detects outlier of human pose coordinates in move motion effectively and rule-based correction method corrects the outlier according to a simple rule. We have shown that the proposed method is effective in various motions with the experiments using 2D golf swing motion data and have shown the possibility of expansion from 2D to 3D coordinates. |
Å°¿öµå(Keyword) |
¿¬ÁÖ ÀÚ¼¼
±×¶ó¿îµå Æ®·ç½º
¹ÝÀÚµ¿ ÁÖ¼®È ¹æ¹ý
ÀΰøÁö´É
ÄÄÇ»ÅÍ ºñÀü
Playing Posture
Ground-Truth
Semi-Automatic Annotation Method
Artificial Intelligence
Computer Vision
»ç¶÷ ÀÚ¼¼ ÃßÁ¤
ÁÂÇ¥ º¸Á¤
µö·¯´×
Human Pose Estimation
Pose Refinement
Deep Learning
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|