Skip to content

IAMonSense: Multi-level Handwriting Classification using Spatio-temporal Information | Embedded Intelligence | RPTU DFKI

License

Notifications You must be signed in to change notification settings

t4f1d/iamonsense

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

IAMonSense: Multi-level Handwriting Classification using Spatio-temporal Information

Updates

06 2023: Paper Published at IJDAR
01 2023: Initial commit

About

This repository is the implementation of Multi-level Handwriting Classification using Spatio-temporal Information, as described in the following paper: IAMonSense: Multi-level Handwriting Classification using Spatio-temporal Information, by A. Mustafid, J. Younas, P. Lukowicz, S. Ahmed.

Link Paper : https://link.springer.com/article/10.1007/s10032-023-00433-y paper-iamonsense

Our paper investigates online handwriting classification which preliminary step for recognition systems and others applications. The problem in this research is, in which level is good for the classification. Each level has benefits and drawbacks. However, there has been little discussion on it and few researchers have addressed the problem because the datasets are also not comparable at all levels. Thus, our research aimed to enhance the datasets and also broaden current knowledge of classification online-handwriting in multi-classes and in multi-levels.

Data Multilevel

Without datasets that contain proper data and information at all levels, it is insufficient to do a multi-level online handwriting classification. Consequently, we need to prepare, enhance, and enrich the datasets to be able to perform multi-level classification. We enhance the existing online-handwriting dataset (SenseThePen, IAM-OnDB, and IAMonDo) to be able to do multi-level online handwriting classification.

Datasets Comparison

The contributions of this research establish a foundation and serve as a baseline for systematic and empirical evaluation of online handwriting classification. We provide a new enhancement dataset for the research community, the dataset called IAMonSense. It can be used for graph models or deep learning models with different data structure. It contains multi-level information from stroke, word, and line levels. This research has highlighted the importance of line level in the classification problem.

References:

[14] J. Younas, M. I. Malik, S. Ahmed, F. Shafait, and P. Lukowicz, "Sense the pen: Classification of online handwritten sequences (text, mathematical expression, plot/graph)", Expert Systems with Applications, vol. 172, p. 114 588, 2021. doi: https://doi.org/10.1016/j.eswa.2021.114588.
[27] E. Indermühle, M. Liwicki, and H. Bunke, "IAMonDo-Database: An Online Handwritten Document Database with Non-Uniform Contents", in Proceedings of the 9th IAPR International Workshop on Document Analysis Systems, ser. DAS ’10, Boston, Massachusetts, USA: Association for Computing Machinery, 2010, pp. 97–104. doi: https://doi.org/10.1145/1815330.1815343.
[28] M. Liwicki and H. Bunke, "IAM-OnDB - an On-Line English Sentence Database Acquired from Handwritten Text on a Whiteboard", in Proceedings of the Eighth International Conference on Document Analysis and Recognition, ser. ICDAR ’05, USA: IEEE Computer Society, 2005, pp. 956–961. doi: https://doi.org/10.1109/ICDAR.2005.132.

Dataset

Download

You can download the dataset from here (SeaFile) or here (GDrive).

Structure

The structure of the IAMonSense dataset,

IAMonSense/
├─ SenseThePen+/
│  └─ line_data/
│     ├─ p1/
│     │  ├─ l_1.csv
│     │  ├─ l_2.csv
│     │  ├─ l_3.csv
│     │  ├─ ...
│     │  └─ l_21.csv
│     ├─ p2/
│     │  ├─ l_1.csv
│     │  ├─ l_2.csv
│     │  ├─ l_3.csv
│     │  ├─ ...
│     │  └─ l_72.csv
│     ├─ .../
│     └─ p20/
│        ├─ l_1.csv
│        ├─ l_2.csv
│        ├─ l_3.csv
│        ├─ ...
│        └─ l_83.csv
│
├─ IAMonDo+/
│  ├─ 001.csv
│  ├─ 001a.csv
│  ├─ 001b.csv
│  ├─ 001c.csv
│  ├─ 001d.csv
│  ├─ 001e.csv
│  ├─ ...
│  └─ 982.csv
│
├─ IAM-OnDB+/
│  └─ line_data/
│     ├─ a01-000/
│     │  ├─ a01-000u-01.csv
│     │  ├─ a01-000u-02.csv
│     │  ├─ a01-000u-03.csv
│     │  ├─ ...
│     │  └─ a01-000u-06.csv
│     ├─ a01-001/
│     │  ├─ a01-001w-01.csv
│     │  ├─ a01-001w-02.csv
│     │  ├─ a01-001w-03.csv
│     │  ├─ ...
│     │  └─ a01-001z-09.csv
│     ├─ .../
│     └─ z01-000/
│        ├─ z01-000-01.csv
│        ├─ z01-000-02.csv
│        ├─ z01-000-03.csv
│        ├─ ...
│        └─ z01-000z-08.csv
│
├─ SenseThePen_train.csv
├─ SenseThePen_val.csv
├─ SenseThePen_test.csv
├─ IAMonDo_train.csv
├─ IAMonDo_val.csv
├─ IAMonDo_test.csv
├─ IAM-OnDB_train.csv
├─ IAM-OnDB_val.csv
└─ IAM-OnDB_test.csv

Statistics

Datasets # of files # of strokes # of words # of lines # of classes
SenseThePen+ 1,595 36,329 12,947 1,595 3
IAMonDo+ 941 356,189 87,924 18,658 3
IAM-OnDB+ 12,190 304,696 64,084 12,190 1

Results

Deep Learning Model

Performance Analysis Deep Learning Model

Graph Model

Performance Analysis Graph Model

Transformer Model

Performance Analysis Transformer Model

State-of-the-art Comparison

State-of-the-art Comparison

References:

[10] J.-Y. Ye, Y.-M. Zhang, Q. Yang, and C.-L. Liu, “Contextual stroke classification in online handwritten documents with edge graph attention networks”, SN Computer Science, vol. 1, no. 3, pp. 1–13, 2020. doi: https://doi.org/10.1007/s42979-020-00177-0.
[14] J. Younas, M. I. Malik, S. Ahmed, F. Shafait, and P. Lukowicz, “Sense the pen: Classification of online handwritten sequences (text, mathematical expression, plot/graph)”, Expert Systems with Applications, vol. 172, p. 114 588, 2021. doi: https://doi.org/10.1016/j.eswa.2021.114588.
[15] J. Younas, S. Fritsch, G. Pirkl, S. Ahmed, M. I. Malik, F. Shafait, and P. Lukowicz, “What Am I Writing: Classification of On-Line Handwritten Sequences.”, in Intelligent Environments (Workshops), ser. Ambient Intelligence and Smart Environments, vol. 23, IOS Press, 2018, pp. 417–426.
[34] I. Degtyarenko, I. Deriuga, A. Grygoriev, S. Polotskyi, V. Melnyk, D. Zakharchuk, and O. Radyvonenko, “Hierarchical Recurrent Neural Network for Handwritten Strokes Classification”, in ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021, pp. 2865–2869. doi: https://doi.org/10.1109/ICASSP39728.2021.9413412.

License

This project is licensed under the MIT License. See LICENSE for more details.

Citations

@article{Mustafid2023, 
    title={IAMonSense: Multi-level handwriting classification using spatiotemporal information}, 
    DOI={10.1007/s10032-023-00433-y}, 
    journal={International Journal on Document Analysis and Recognition (IJDAR)}, 
    author={Mustafid, Ahmad and Younas, Junaid and Lukowicz, Paul and Ahmed, Sheraz}, 
    year={2023}
} 

Acknowledgements

We would like to thank the following people for their support: Siti Helmiyah, M. Murah Pamuji, Boby Gunarso, Noor Titan Putri Hartono, Sukma Dyah Aini. Special thanks to RPTU and DFKI.

About

IAMonSense: Multi-level Handwriting Classification using Spatio-temporal Information | Embedded Intelligence | RPTU DFKI

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published