TY - GEN
T1 - Image Collation
T2 - 16th International Conference on Document Analysis and Recognition, ICDAR 2021
AU - Kaoua, Ryad
AU - Shen, Xi
AU - Durr, Alexandra
AU - Lazaris, Stavros
AU - Picard, David
AU - Aubry, Mathieu
N1 - Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
PY - 2021/1/1
Y1 - 2021/1/1
N2 - Illustrations are an essential transmission instrument. For an historian, the first step in studying their evolution in a corpus of similar manuscripts is to identify which ones correspond to each other. This image collation task is daunting for manuscripts separated by many lost copies, spreading over centuries, which might have been completely re-organized and greatly modified to adapt to novel knowledge or belief and include hundreds of illustrations. Our contributions in this paper are threefold. First, we introduce the task of illustration collation and a large annotated public dataset to evaluate solutions, including 6 manuscripts of 2 different texts with more than 2 000 illustrations and 1 200 annotated correspondences. Second, we analyze state of the art similarity measures for this task and show that they succeed in simple cases but struggle for large manuscripts when the illustrations have undergone very significant changes and are discriminated only by fine details. Finally, we show clear evidence that significant performance boosts can be expected by exploiting cycle-consistent correspondences. Our code and data are available on http://imagine.enpc.fr/~shenx/ImageCollation.
AB - Illustrations are an essential transmission instrument. For an historian, the first step in studying their evolution in a corpus of similar manuscripts is to identify which ones correspond to each other. This image collation task is daunting for manuscripts separated by many lost copies, spreading over centuries, which might have been completely re-organized and greatly modified to adapt to novel knowledge or belief and include hundreds of illustrations. Our contributions in this paper are threefold. First, we introduce the task of illustration collation and a large annotated public dataset to evaluate solutions, including 6 manuscripts of 2 different texts with more than 2 000 illustrations and 1 200 annotated correspondences. Second, we analyze state of the art similarity measures for this task and show that they succeed in simple cases but struggle for large manuscripts when the illustrations have undergone very significant changes and are discriminated only by fine details. Finally, we show clear evidence that significant performance boosts can be expected by exploiting cycle-consistent correspondences. Our code and data are available on http://imagine.enpc.fr/~shenx/ImageCollation.
UR - https://www.scopus.com/pages/publications/85115288209
U2 - 10.1007/978-3-030-86337-1_24
DO - 10.1007/978-3-030-86337-1_24
M3 - Conference contribution
AN - SCOPUS:85115288209
SN - 9783030863364
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 351
EP - 366
BT - Document Analysis and Recognition - ICDAR 2021 - 16th International Conference, Proceedings
A2 - Lladós, Josep
A2 - Lopresti, Daniel
A2 - Uchida, Seiichi
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 5 September 2021 through 10 September 2021
ER -