Á¤º¸Ã³¸®ÇÐȸ ³í¹®Áö ¼ÒÇÁÆ®¿þ¾î ¹× µ¥ÀÌÅÍ °øÇÐ
Current Result Document : 1 / 5
´ÙÀ½°Ç
ÇѱÛÁ¦¸ñ(Korean Title) |
´ëÇѹα¹ Á¤ºÎÀÇ Äڷγª 19 ºê¸®ÇÎÀ» ±â¹ÝÀ¸·Î ±¸ÃàµÈ ¼ö¾î µ¥ÀÌÅͼ ¿¬±¸ |
¿µ¹®Á¦¸ñ(English Title) |
Sign Language Dataset Built from S. Korean Government Briefing on COVID-19 |
ÀúÀÚ(Author) |
Hohyun Sim
Horyeol Sung
Seungjae Lee
Hyeonjoong Cho
½ÉÈ£Çö
¼ºÈ£·Ä
À̽ÂÀç
Á¶ÇöÁß
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 11 NO. 08 PP. 0325 ~ 0330 (2022. 08) |
Çѱ۳»¿ë (Korean Abstract) |
º» ³í¹®Àº Çѱ¹ ¼ö¾î¿¡ ´ëÇÏ¿© ¼ö¾î ÀνÄ, ¼ö¾î ¹ø¿ª, ¼ö¾î ¿µ»ó ½ÃºÐÇÒ°ú °°Àº ¼ö¾î¿¡ °üÇÑ µö·¯´× ¿¬±¸¸¦ À§ÇÑ µ¥ÀÌÅͼÂÀÇ ¼öÁý ¹× ½ÇÇèÀ» ÁøÇàÇÏ¿´´Ù. ¼ö¾î ¿¬±¸¸¦ À§ÇÑ ¾î·Á¿òÀº 2°¡Áö·Î º¼ ¼ö ÀÖ´Ù. ù°, ¼ÕÀÇ ¿òÁ÷ÀÓ°ú ¼ÕÀÇ ¹æÇâ, Ç¥Á¤ µîÀÇ Á¾ÇÕÀûÀÎ Á¤º¸¸¦ °¡Áö´Â ¼ö¾îÀÇ Æ¯¼º¿¡ µû¸¥ ÀνÄÀÇ ¾î·Á¿òÀÌ ÀÖ´Ù. µÑ°, µö·¯´× ¿¬±¸¸¦ ÁøÇàÇϱâ À§ÇÑ ÇнÀµ¥ÀÌÅÍÀÇ Àý´ëÀû ºÎÀçÀÌ´Ù. ÇöÀç ¾Ë·ÁÁø ¹®Àå ´ÜÀ§ÀÇ Çѱ¹ ¼ö¾î µ¥ÀÌÅͼÂÀº KETI µ¥ÀÌÅͼÂÀÌ À¯ÀÏÇÏ´Ù. ÇØ¿ÜÀÇ ¼ö¾î µö·¯´× ¿¬±¸¸¦ À§ÇÑ µ¥ÀÌÅͼÂÀº Isolated ¼ö¾î¿Í Continuous ¼ö¾î µÎ °¡Áö·Î ºÐ·ùµÇ¾î ¼öÁýµÇ¸ç ½Ã°£ÀÌ Áö³¯¼ö·Ï ´õ ¸¹Àº ¾çÀÇ ¼ö¾î µ¥ÀÌÅÍ°¡ ¼öÁýµÇ°í ÀÖ´Ù. ÇÏÁö¸¸ ÀÌ·¯ÇÑ ÇØ¿ÜÀÇ ¼ö¾î µ¥ÀÌÅͼµµ ¹æ´ëÇÑ µ¥ÀÌÅͼÂÀ» ÇÊ¿ä·Î ÇÏ´Â µö·¯´× ¿¬±¸¸¦ À§Çؼ´Â ºÎÁ·ÇÑ »óȲÀÌ´Ù. º» ¿¬±¸¿¡¼´Â Çѱ¹ ¼ö¾î µö·¯´× ¿¬±¸¸¦ ÁøÇàÇϱâ À§ÇÑ ´ë±Ô¸ðÀÇ Çѱ¹¾î-¼ö¾î µ¥ÀÌÅͼÂÀ» ¼öÁýÀ» ½ÃµµÇÏ¿´À¸¸ç º£À̽º¶óÀÎ ¸ðµ¨À» ÀÌ¿ëÇÏ¿© ¼ö¾î ¹ø¿ª ¸ðµ¨ÀÇ ¼º´É Æò°¡ ½ÇÇèÀ» ÁøÇàÇÏ¿´´Ù. º» ³í¹®À» À§ÇØ ¼öÁýµÈ µ¥ÀÌÅͼÂÀº ÃÑ 11,402°³ÀÇ ¿µ»ó°ú ÅؽºÆ®·Î ±¸¼ºµÇ¾ú´Ù. À̸¦ ÀÌ¿ëÇÏ¿© ÇнÀÀ» ÁøÇàÇÒ º£À̽º¶óÀÎ ¸ðµ¨·Î´Â ¼ö¾î ¹ø¿ª ºÐ¾ß¿¡¼ SOTAÀÇ ¼º´ÉÀ» °¡Áö°í ÀÖ´Â TSPNet ¸ðµ¨À» ÀÌ¿ëÇÏ¿´´Ù. º» ³í¹®ÀÇ ½ÇÇè¿¡¼ ¼öÁýµÈ µ¥ÀÌÅͼ¿¡ ´ëÇÑ Æ¯¼ºÀ» Á¤·®ÀûÀ¸·Î º¸ÀÌ°í, º£À̽º¶óÀÎ ¸ðµ¨ÀÇ ½ÇÇè °á°ú·Î´Â BLEU-4 score 3.63À» º¸¿´´Ù. ¶ÇÇÑ, ÇâÈÄ ¿¬±¸¿¡¼ º¸´Ù Á¤È®ÇÏ°Ô µ¥ÀÌÅͼÂÀ» ¼öÁýÇÒ ¼ö ÀÖµµ·Ï, Çѱ¹¾î–¼ö¾î µ¥ÀÌÅͼ ¼öÁý¿¡ ÀÖ¾î¼ °í·ÁÇÒ Á¡À» Æò°¡ °á°ú¿¡ ´ëÇÑ °íÂû·Î Á¦½ÃÇÑ´Ù. |
¿µ¹®³»¿ë (English Abstract) |
This paper conducts the collection and experiment of datasets for deep learning research on sign language such as sign language recognition, sign language translation, and sign language segmentation for Korean sign language. There exist difficulties for deep learning research of sign language. First, it is difficult to recognize sign languages since they contain multiple modalities including hand movements, hand directions, and facial expressions. Second, it is the absence of training data to conduct deep learning research. Currently, KETI dataset is the only known dataset for Korean sign language for deep learning. Sign language datasets for deep learning research are classified into two categories: Isolated sign language and Continuous sign language. Although several foreign sign language datasets have been collected over time. they are also insufficient for deep learning research of sign language. Therefore, we attempted to collect a large-scale Korean sign language dataset and evaluate it using a baseline model named TSPNet which has the performance of SOTA in the field of sign language translation. The collected dataset consists of a total of 11,402 image and text. Our experimental result with the baseline model using the dataset shows BLEU-4 score 3.63, which would be used as a basic performance of a baseline model for Korean sign language dataset. We hope that our experience of collecting Korean sign language dataset helps facilitate further research directions on Korean sign language. |
Å°¿öµå(Keyword) |
Sign Language Recognition
Sign Language Translation
Sign Language Segmentation
Sign Language Dataset
Deep Learning
¼ö¾î ÀνÄ
¼ö¾î ¹ø¿ª
¼ö¾î ¿µ»ó ½ÃºÐÇÒ
¼ö¾î µ¥ÀÌÅͼÂ
µö·¯´×
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|