Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
ÀÓº£µðµå Àåºñ ȯ°æ¿¡¼ °æ·®ÈµÈ °´Ã¼ ŽÁö µö·¯´× ¸ðµ¨ ºñ±³ |
¿µ¹®Á¦¸ñ(English Title) |
A Comparison of Lightweight Object Detection Deep Learning Model in Embedded Environment |
ÀúÀÚ(Author) |
±èÁö³ª
Á¤ÀºÁö
ÀüÁ¦¼º
À̼ö¾È
Gina Kim
Eunji Jeong
Jaesung Jun
Suan Lee
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 45 NO. 01 PP. 2165 ~ 2168 (2022. 06) |
Çѱ۳»¿ë (Korean Abstract) |
|
¿µ¹®³»¿ë (English Abstract) |
Object detection, one of the representative technologies of artificial intelligence, is used in various fields such as smart cities and autonomous driving and is becoming increasingly important. Recently, in order to use artificial intelligence technology in the Edge environment, many researchers are studying how to reduce the weight of the model. In this paper, we compare YOLOv5, EfficientDet, SSD MobileNetV1, and spaghettinet models in terms of inference time. For comparative experiments, embedded devices used Google Coral Dev Board and ASUS Tinker Edge T. |
Å°¿öµå(Keyword) |
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|