Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
ºñµð¿À ¼¼±×¸ÕÆ® ´ÜÀ§ÀÇ ºÎºÐ º¹»ç °ËÃâÀ» À§ÇÑ CNN ±â¹Ý ÇÁ·¹ÀÓ Æ¯Â¡ º¤ÅÍ À¶ÇÕ ¹æ¹ý |
¿µ¹®Á¦¸ñ(English Title) |
A Fusion of CNN-based Frame Vector for Segment-level Video Partial Copy Detection |
ÀúÀÚ(Author) |
ÃÖÁ¤È¯
ÃÖÁö¿ø
·ù´ö»ê
±è¼øÅÂ
Jeongwhan Choi
Jiwon Choi
Duksan Ryu
Suntae Kim
Á¤¹Î¼ö
³¶Á¾È£
Minsoo Jeong
Jongho Nang
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 48 NO. 01 PP. 0043 ~ 0050 (2021. 01) |
Çѱ۳»¿ë (Korean Abstract) |
ÃÖ±Ù À¯Æ©ºê³ª ÀνºÅ¸±×·¥°ú °°Àº ÄÜÅÙÃ÷ Ç÷§ÆûÀ» ÁÖÃàÀ¸·Î ¹Ìµð¾î¿¡ ´ëÇÑ ¼ö¿ä°¡ ±Þ¼ÓÇÏ°Ô Áõ°¡ÇÏ°í ÀÖ´Ù. ÀÌ¿¡ µû¶ó ÀúÀÛ±Ç º¸È£³ª ºÒ¹ý ÄÜÅÙÃ÷ÀÇ À¯Æ÷¿Í °°Àº ¹®Á¦µéÀÌ ¹ß»ýÇÏ°í ÀÖ´Ù. ÀÌ·¯ÇÑ ¹®Á¦¸¦ ÇØ°áÇϱâ À§ÇØ ³»¿ë¿¡ ±â¹ÝÇÑ °íÀ¯ÀÇ ½Äº°ÀÚ¸¦ ÃßÃâÇÏ´Â ¹æ¹ýµéÀÌ Á¦¾ÈµÇ¾úÁö¸¸ ±âÁ¸ÀÇ ¿¬±¸µéÀº ¹Ì¸® Á¤ÇØÁø º¯Çü¿¡ ´ëÇÏ¿© °í¾ÈµÇ¾ú±â ¶§¹®¿¡ ½ÇÁ¦ ºñµð¿À¿¡¼´Â °ËÃâ¿¡ ½ÇÆÐÇÏ¿´´Ù. º» ³í¹®¿¡¼´Â ½ÇÁ¦ À¯ÅëµÇ´Â ºñµð¿ÀÀÇ ´Ù¾çÇÑ º¯Çü¿¡ °ÀÎÇÑ ºÎºÐ º¹»ç °ËÃâÀ» À§ÇØ ÇÁ·¹ÀÓ Á¤º¸¸¦ À¶ÇÕÇÑ µö·¯´× ±â¹ÝÀÇ ¼¼±×¸ÕÆ® Fingerprint¸¦ Á¦¾ÈÇÑ´Ù. TIRI¸¦ ÀÌ¿ëÇÑ µ¥ÀÌÅÍ ¼öÁØÀÇ À¶ÇÕ ¹æ¹ý°ú Ç®¸µÀ» ÀÌ¿ëÇÑ Æ¯Â¡ º¤ÅÍ ¼öÁØÀÇ À¶ÇÕ ¹æ¹ýÀ¸·Î ÃßÃâÇÑ Fingerprint¸¦ Triplet loss¸¦ ÀÌ¿ëÇÏ¿© ÇнÀÇÏ°í °ËÃ⠽ýºÅÛÀ» ¼³°èÇÏ¿© ¼º´ÉÀ» ºÐ¼®ÇÑ´Ù. º» ³í¹®ÀÇ ½ÇÇèÀº À¯Æ©ºê¸¦ ±â¹ÝÀ¸·Î ¼öÁýÇÑ µ¥ÀÌÅͼÂÀÎ VCDB¸¦ ÀÌ¿ëÇÏ¿´À¸¸ç 5ÃÊ µ¿¾È »ùÇøµÇÑ ÇÁ·¹ÀÓ Æ¯Â¡ º¤Å͸¦ Max Ç®¸µÀ¸·Î À¶ÇÕÇÏ¿© 66%ÀÇ ¼º´ÉÀ» ¾ò¾ú´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
Recently, the demand for media has grown rapidly, led by multimedia content platforms such as YouTube and Instagram. As a result, problems such as copyright protection and the spread of illegal content have arisen. To solve these problems, studies have been proposed to extract unique identifiers based on the content. However, existing studies were designed for simulated transformation and failed to detect whether the copied videos were actually shared. In this paper, we proposed a deep learning-based segment fingerprint that fused frame information for partial copy detection that was robust for various variations in the actually shared video. We used TIRI for data-level fusion and Pooling for feature-level fusion. We also designed a detection system with a segment fingerprint that was trained with Triplet loss. We evaluated the performance with VCDB, a dataset collected based on YouTube, and obtained 66% performance by fusing frame features sampled for 5 seconds with Max pooling for detecting video partial-copy problems.
|
Å°¿öµå(Keyword) |
ÄÁÇDZԷ¹ÀÌ¼Ç ¹ö±× ¸®Æ÷Æ®
¼±ÇüÆǺ°ºÐ¼®
Â÷¿øÃà¼Ò
Ŭ·¡½º ºÒ±ÕÇü
»ùÇøµ
configuration bug report
linear discriminant analysis
dimensionality reduction
class imbalance
sampling
ºñµð¿À
ºñµð¿À º¹»ç °ËÃâ
ºñµð¿À ºÎºÐ º¹»ç
Ư¡ À¶ÇÕ
CNN
µö·¯´×
CNN
video analysis
copy detection
partial-copy detection
feature fusion
deep learn
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|