KCC 2021
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
TrackNet-V2¸¦ »ç¿ëÇÑ °ñÇÁ °ø ÃßÀû |
¿µ¹®Á¦¸ñ(English Title) |
Tracking of Golf Ball using TrackNetV2 |
ÀúÀÚ(Author) |
¹ÙÀ̵ð¾Æ ·±ÀÚÀÌ
ÀÌ»ó¿õ
Ranjai Baidya
Sang-Woong Lee
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 48 NO. 01 PP. 0659 ~ 0660 (2021. 06) |
Çѱ۳»¿ë (Korean Abstract) |
|
¿µ¹®³»¿ë (English Abstract) |
Knowing the position of a golf ball during its flight is important for both the player as well as the viewers. To know the exact position of the ball is a difficult task as it moves in a very high speed. Using readily available tools like launch monitors is an option but, they are expensive and inaccessible. Here we have used a deep neural network, TrackNetV2 to track a golf ball when it is in motion. Additionally, the direction of the golf ball has been brought into consideration to get better results as compared to using only the TrackNet-v2 model. The used system would take a mono-stereo video as input and output a video with golf ball labelled over it. The trained model was applied on a series of golf shot videos and reasonable outputs were obtained |
Å°¿öµå(Keyword) |
golf ball
golf club
TrackNetV2
Club
Tracking
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|