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

»çÀÌÆ®¸Ê

Loading..

Please wait....

Çмú´ëȸ ÇÁ·Î½Ãµù

Ȩ Ȩ > ¿¬±¸¹®Çå > Çмú´ëȸ ÇÁ·Î½Ãµù > IPIU (¿µ»óó¸® ¹× ÀÌÇØ¿¡ °üÇÑ ¿öÅ©¼¥) > IPIU 1997 (Á¦9ȸ ¿µ»óó¸® ¹× ÀÌÇØ¿¡ °üÇÑ ¿öÅ©¼¥)

IPIU 1997 (Á¦9ȸ ¿µ»óó¸® ¹× ÀÌÇØ¿¡ °üÇÑ ¿öÅ©¼¥)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) Projection ProfileÀ» ÀÌ¿ëÇÑ »õ·Î¿î ÀÚµ¿ ¹®¼­¿µ»óÀÇ ¿µ¿ªºÐ¸® ¹× ºÐ·ù ¾Ë°í¸®µë
¿µ¹®Á¦¸ñ(English Title) A new algorithm for automatic blocking and classification of a document image using projection profile
ÀúÀÚ(Author) Á¶Çö¸ñ   À̰湫   ÃÖ¿µ¿ì   Hyun Mok Cho   Kyoung Mu Lee   Yeong Woo Choi  
¿ø¹®¼ö·Ïó(Citation) VOL 09 NO. 01 PP. 0136 ~ 0140 (1997. 01)
Çѱ۳»¿ë
(Korean Abstract)
¿µ¹®³»¿ë
(English Abstract)
In this research we propose a new block segmentation and classification algorithm for a binary document image based on a recursive X-Y cut technique of projection profiles of the image. By using horizontal projection profile of the reduced image in horizontal direction and vertical projection profile of the reduced image in vertical direction alternatively, the proposed algorithm can enhance the computational efficiency as well as the cutting resolution in block segmentation, and thus achieves fast and accurate blocking results. Each segmented document block is then classified into text, halftone and diagram according to features such as horizontal projection profile pattern and B/W (black and white) ratio of the block. In order to extract skewed text region robustly and accurately, we use a sliced window profile matching technique, in which several local Y-projection profiles are compared to ensure the text region. Simulation results are shown to demonstrate the performance of the proposed algorithm on many scanned document images.
Å°¿öµå(Keyword)
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå