Open Images V5 Text Annotation and Yet Another Mask Text Spotter

Ilya Krylov, Sergei Nosov, Vladislav Sovrasov
Proceedings of The 13th Asian Conference on Machine Learning, PMLR 157:379-389, 2021.

Abstract

A large scale human-labeled dataset plays an important role in creating high quality deep learning models. In this paper we present text annotation for Open Images V5 dataset. To our knowledge it is the largest among publicly available manually created text annotations. Having this annotation we trained a simple Mask-RCNN-based network, referred as Yet Another Mask Text Spotter (YAMTS), which achieves competitive performance or even outperforms current state-of-the-art approaches in some cases on ICDAR 2013, ICDAR 2015 and {Total-Text} datasets. Code for text spotting model available online at: \url{https://github.com/openvinotoolkit/training_extensions}. The model can be exported to OpenVINO{\texttrademark}-format and run on Intel{\textregistered} CPUs.

Cite this Paper

BibTeX
@InProceedings{pmlr-v157-krylov21a, title = {Open Images V5 Text Annotation and Yet Another Mask Text Spotter}, author = {Krylov, Ilya and Nosov, Sergei and Sovrasov, Vladislav}, booktitle = {Proceedings of The 13th Asian Conference on Machine Learning}, pages = {379--389}, year = {2021}, editor = {Balasubramanian, Vineeth N. and Tsang, Ivor}, volume = {157}, series = {Proceedings of Machine Learning Research}, month = {17--19 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v157/krylov21a/krylov21a.pdf}, url = {https://proceedings.mlr.press/v157/krylov21a.html}, abstract = {A large scale human-labeled dataset plays an important role in creating high quality deep learning models. In this paper we present text annotation for Open Images V5 dataset. To our knowledge it is the largest among publicly available manually created text annotations. Having this annotation we trained a simple Mask-RCNN-based network, referred as Yet Another Mask Text Spotter (YAMTS), which achieves competitive performance or even outperforms current state-of-the-art approaches in some cases on ICDAR 2013, ICDAR 2015 and {Total-Text} datasets. Code for text spotting model available online at: \url{https://github.com/openvinotoolkit/training_extensions}. The model can be exported to OpenVINO{\texttrademark}-format and run on Intel{\textregistered} CPUs.} }
Endnote
%0 Conference Paper %T Open Images V5 Text Annotation and Yet Another Mask Text Spotter %A Ilya Krylov %A Sergei Nosov %A Vladislav Sovrasov %B Proceedings of The 13th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Vineeth N. Balasubramanian %E Ivor Tsang %F pmlr-v157-krylov21a %I PMLR %P 379--389 %U https://proceedings.mlr.press/v157/krylov21a.html %V 157 %X A large scale human-labeled dataset plays an important role in creating high quality deep learning models. In this paper we present text annotation for Open Images V5 dataset. To our knowledge it is the largest among publicly available manually created text annotations. Having this annotation we trained a simple Mask-RCNN-based network, referred as Yet Another Mask Text Spotter (YAMTS), which achieves competitive performance or even outperforms current state-of-the-art approaches in some cases on ICDAR 2013, ICDAR 2015 and {Total-Text} datasets. Code for text spotting model available online at: \url{https://github.com/openvinotoolkit/training_extensions}. The model can be exported to OpenVINO{\texttrademark}-format and run on Intel{\textregistered} CPUs.
APA
Krylov, I., Nosov, S. & Sovrasov, V.. (2021). Open Images V5 Text Annotation and Yet Another Mask Text Spotter. Proceedings of The 13th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 157:379-389 Available from https://proceedings.mlr.press/v157/krylov21a.html.

Related Material