• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ÇÐȸÁö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ÇÐȸÁö > µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

µ¥ÀÌÅͺ£À̽º ¿¬±¸È¸Áö(SIGDB)

Current Result Document : 10 / 10

ÇѱÛÁ¦¸ñ(Korean Title) À̹ÌÁö ³»ÀÇ ÅؽºÆ® µ¥ÀÌÅÍ ÀÎ½Ä Á¤È®µµ Çâ»óÀ» À§ÇÑ ¸ÖƼ ¸ð´Þ À̹ÌÁö ó¸® ÇÁ·Î¼¼½º
¿µ¹®Á¦¸ñ(English Title) Multi-modal Image Processing for Improving Recognition Accuracy of Text Data in Images
ÀúÀÚ(Author) ¹ÚÁ¤Àº   ÁÖ°æµ·   ±èö¿¬   Jungeun Park   Gyeongdon Joo   Chulyun Kim  
¿ø¹®¼ö·Ïó(Citation) VOL 34 NO. 03 PP. 0148 ~ 0158 (2018. 12)
Çѱ۳»¿ë
(Korean Abstract)
±¤ÇÐ ¹®ÀÚ ÀνÄ(OCR)Àº ÅؽºÆ®¸¦ Æ÷ÇÔÇÑ À̹ÌÁö¿¡¼­ ÅؽºÆ® ¿µ¿ªÀ» ÀνÄÇÏ°í À̷κÎÅÍ ÅؽºÆ®¸¦ ÃßÃâÇÏ´Â ±â¼úÀÌ´Ù. Àüü ÅؽºÆ® µ¥ÀÌÅÍ Áß »ó´çÈ÷ ¸¹Àº ÅؽºÆ® Á¤º¸°¡ À̹ÌÁö¿¡ Æ÷ÇԵǾî Àֱ⠶§¹®¿¡ OCRÀº µ¥ÀÌÅÍ ºÐ¼® ºÐ¾ß¿¡ ÀÖ¾î Áß¿äÇÑ Àüó¸® ´Ü°è¸¦ ´ã´çÇÑ´Ù. ´ëºÎºÐÀÇ OCR ¿£ÁøÀÌ, Èò ¹ÙÅÁÀÇ °ËÁ¤ ±Û¾¾ÀÇ ´Ü¼øÇÑ ÇüŸ¦ °¡Áø À̹ÌÁö¿Í °°Àº, ÅؽºÆ®¿Í ¹è°æÀÇ ±¸ºÐÀÌ ¶Ñ·ÇÇÑ Àú º¹Àâµµ À̹ÌÁö¿¡ ´ëÇؼ­´Â ³ôÀº ÀνķüÀ» º¸ÀÌ´Â ¹Ý¸é, ÅؽºÆ®¿Í ¹è°æÀÇ ±¸ºÐÀÌ ¶Ñ·ÇÇÏÁö ¾ÊÀº °í º¹Àâµµ À̹ÌÁö¿¡ ´ëÇؼ­´Â ÀúÁ¶ÇÑ ÀνķüÀ» º¸À̱⠶§¹®¿¡, Àνķü °³¼±À» À§ÇØ ÀÔ·Â À̹ÌÁö¸¦ OCR ¿£ÁøÀÌ Ã³¸®Çϱ⠿ëÀÌÇÑ À̹ÌÁö·Î º¯ÇüÇÏ´Â Àüó¸® ÀÛ¾÷ÀÌ ÇÊ¿äÇÏ°Ô µÈ´Ù. µû¶ó¼­ º» ³í¹®¿¡¼­´Â OCR ¿£ÁøÀÇ Á¤È®¼º Áõ´ë¸¦ À§ÇØ ÅؽºÆ® ¶óÀκ°·Î À̹ÌÁö¸¦ ºÐ¸®ÇÏ°í, ¿µ»óó¸® ±â¹ý ±â¹ÝÀÇ CLAHE ¸ðµâ°ú Two-step ¸ðµâÀ» º´·ÄÀûÀ¸·Î ¼öÇàÇÏ¿© ÅؽºÆ®¿Í ¹è°æ ¿µ¿ªÀ» È¿À²ÀûÀ¸·Î ºÐ¸®ÇÑ ÈÄ ÅؽºÆ®¸¦ ÀνÄÇÑ´Ù. À̾ µÎ ¸ðµâÀÇ °á°ú ÅؽºÆ®¿¡ ´ëÇÏ¿© N-gram¹æ¹ý°ú Hunspell »çÀüÀ» °áÇÕÇÑ ¾Ë°í¸®ÁòÀ¸·Î ÀνķüÀ» ºñ±³ÇÏ¿© °¡Àå ³ôÀº ÀνķüÀÇ °á°ú ÅؽºÆ®¸¦ ÃÖÁ¾ °á°ú¹°·Î ¼±Á¤ÇÏ´Â ¹æ¹ý·ÐÀ» Á¦¾ÈÇÑ´Ù. ´ëÇ¥ÀûÀÎ OCR ¿£ÁøÀÎ Tesseract¿Í Abbyy¿ÍÀÇ ´Ù¾çÇÑ ºñ±³ ½ÇÇèÀ» ÅëÇØ º» ¿¬±¸¿¡¼­ Á¦¾ÈÇÏ´Â ¸ðµâÀÌ º¹ÀâÇÑ ¹è°æÀ» °¡Áø À̹ÌÁö¿¡¼­ °¡Àå Á¤È®ÇÑ ÅؽºÆ® ÀνķüÀ» º¸ÀÓÀ» º¸¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
The optical character recognition (OCR) is a technique to extract and recognize texts from images. It is an important preprocessing step in data analysis since most actual text information is embedded in images. Many OCR engines have high recognition accuracy for images where texts are clearly separable from background, such as white background and black lettering. However, they have low recognition accuracy for images where texts are not easily separable from complex background. To improve this low accuracy problem with complex images, it is necessary to transform the input image to make texts more noticeable. In this paper, we propose a method to segment an input image into text lines to enable OCR engines to recognize each line more efficiently, and to determine the final output by comparing the recognition rates of CLAHE module and Two-step module which distinguish texts from background regions based on image processing techniques. Through thorough experiments comparing with well-known OCR engines, Tesseract and Abbyy, we show that our proposed method have the best recognition accuracy with complex background images.
Å°¿öµå(Keyword) µ¥ÀÌÅÍ ÃßÃâ   ±¤Çй®ÀÚÀνĠ  À̹ÌÁö 󸮠  Data Extraction   OCR   Image Processing  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå