TY - GEN
T1 - Simultaneous mixed vertical and horizontal handwritten Japanese character line detection
AU - Kimura, Tomotaka
AU - Premachandra, Chinthaka
AU - Kawanaka, Hiroharu
PY - 2016
Y1 - 2016
N2 - Teachers consume considerable time and their energy in process of marking examination sheets. To reduce this burden on teachers, we have been developing an automatic marking system. To mark examination sheets automatically, handwritten character lines are extracted, and then the characters on those lines are recognized. In this paper, we discuss how character lines are extracted from Japanese handwritten examination sheets without ruled lines. Japanese characters can be written vertically and horizontally, so examination sheets written in Japanese are consisted of mixed vertical and horizontal (MVH) character lines. Conventional character line extraction methods cannot deal with MVH lines, because they have been developed to consider only horizontal character lines. This paper focuses on the simultaneous detection of MVH character lines. The result of experiments using appropriate examination sheet images shows that our method can detect MVH character lines effectively.
AB - Teachers consume considerable time and their energy in process of marking examination sheets. To reduce this burden on teachers, we have been developing an automatic marking system. To mark examination sheets automatically, handwritten character lines are extracted, and then the characters on those lines are recognized. In this paper, we discuss how character lines are extracted from Japanese handwritten examination sheets without ruled lines. Japanese characters can be written vertically and horizontally, so examination sheets written in Japanese are consisted of mixed vertical and horizontal (MVH) character lines. Conventional character line extraction methods cannot deal with MVH lines, because they have been developed to consider only horizontal character lines. This paper focuses on the simultaneous detection of MVH character lines. The result of experiments using appropriate examination sheet images shows that our method can detect MVH character lines effectively.
UR - http://www.scopus.com/inward/record.url?scp=84989884549&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84989884549&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-46418-3_50
DO - 10.1007/978-3-319-46418-3_50
M3 - Conference contribution
AN - SCOPUS:84989884549
SN - 9783319464176
VL - 9972 LNCS
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 564
EP - 572
BT - Computer Vision and Graphics - International Conference, ICCVG 2016, Proceedings
PB - Springer Verlag
T2 - International Conference on Computer Vision and Graphics, ICCVG 2016
Y2 - 19 September 2016 through 21 September 2016
ER -