Download

COCO_TS annotations: COCO_TS_labels.zip [19.1 MB]
The images are the 2014 train images of MSCOCO. They can be downloaded separately at the MSCOCO website.
By downloading the annotations, you agree to our Terms of Use.

If you use this dataset for your research, please cite the following paper:


Bonechi, S., Andreini, P., Bianchini, M., & Scarselli, F. (2019, September). COCO_TS Dataset: Pixel–Level Annotations Based on Weak Supervision for Scene Text Segmentation. In International Conference on Artificial Neural Networks (pp. 238-250). Springer, Cham.


MLT_S annotations: MLT_S_labels.zip [102.0 MB]
The MLT images can be downloaded separately at the MLT dataset download page.
By downloading the annotations, you agree to our Terms of Use.

If you use this dataset for your research, please cite the following paper:


Bonechi, S., Andreini, P., Bianchini, M., & Scarselli, F. (2019). Weak Supervision for Generating Pixel-Level Annotations in Scene Text Segmentation. Pattern Recognition Letters, 2020, ISSN 0167-8655, https://doi.org/10.1016/j.patrec.2020.06.023.


Incidental Scene Text Segmentation annotations: IST_Segmentation_labels.zip [4.3 MB]
The Incidental Scene Text images can be downloaded separately at the Incidental Scene Text dataset download page.
By downloading the annotations, you agree to our Terms of Use.

If you use this dataset for your research, please cite the following paper:


Bonechi, S., Andreini, P., Bianchini, M., & Scarselli, F. (2019). Weak Supervision for Generating Pixel-Level Annotations in Scene Text Segmentation. Pattern Recognition Letters, 2020, ISSN 0167-8655, https://doi.org/10.1016/j.patrec.2020.06.023.


Annotations Description


The semantic label maps are saved as .png images with the same name as the original images of COCO-Text, MLT and Incidental Scene Text datasets.
Pixel values of the provided annotation are defined as follows:

Term of Use

The annotations in these datasets are licensed under a Creative Commons Attribution 4.0 License.