It is quite a challenging task to make a prefect bounding box for a word by hand. This paper deals with this problem, but the algorithm can also be used in the word spotter for finding a perfectly fitting box of the found word.
The image shows how the user has marked the red box, while the algorithm finds the “perfectly” fitting green box. This algorithm makes annotation much easier.
- On-the-fly Historical Text Annotation. Vats, E., Hast, A. In Proceedings of the 14th IAPR International Conference on Document Analysis and Recognition (ICDAR), pp. 10-14, 2017.
@inproceedings{DBLP:conf/icdar/VatsH17, author = {Ekta Vats and Anders Hast}, title = {On-the-fly Historical Handwritten Text Annotation}, booktitle = {1st International Workshop on Human-Document Interaction, 14th {IAPR} International Conference on Document Analysis and Recognition, HDI@ICDAR 2017, Kyoto, Japan, November 9-15, 2017}, pages = {10--14}, year = {2017}, crossref = {DBLP:conf/icdar/2017hdi}, url = {https://doi.org/10.1109/ICDAR.2017.374}, doi = {10.1109/ICDAR.2017.374}}
This image shows how the word spotter finds the bounding box for the word “Firenze”, written by Galileo Galilei, even though the background contains some bleed-through. See the Clavius on the Web Project. Note that the word spotter works on an image where the background has been removed.
Anders Hast 