Detecting Handwritten Line for TikZ Code Generating
Keywords:hough transform, convolutional neural networks, line detection
This work is devoted to the recognition of straight lines in geometric drawings created by hand with the help of neural networks. The LaTeX language is usually used for the design of articles, which takes more time, especially when writing formulas or constructing geometric drawings, unlike writing by hand. Automating the construction of drawings will make it possible to speed up the process of writing articles. The paper considers the recognition of straight lines as the most popular element of geometric drawings. The Hough transform for straight lines detection and its disadvantages are considered. The use of convolutional neural networks for this task is proposed as the best tool for working with images. To train the model, a dataset was created with handwritten lines and lines constructed in a graphics editor. The results of the neural network are given.
WANG Zelun and LIU Jyh-Charn, “Translating math formula images to LaTeX sequences using deep neural networks with sequence-level training,” International Journal on Document Analysis and Recognition (IJDAR), 2021, 24.1–2: 63–75. https://doi.org/10.1007/s10032-020-00360-2
SPRINGSTEIN Matthias, MÜLLER-BUDACK Eric, and EWERTH Ralph, “Unsupervised training data generation of handwritten formulas using generative adversarial networks with self-attention,” In: Proceedings of the 2021 Workshop on Multi-Modal Pre-Training for Multimedia Understanding, 2021, pp. 46–54. https://doi.org/10.1145/3463945.3469059
BIAN Xiaohang, et al., “Handwritten mathematical expression recognition via attention aggregation based bi-directional mutual learning,” In: Proceedings of the AAAI Conference on Artificial Intelligence, 2022, pp. 113–121. https://doi.org/10.1609/aaai.v36i1.19885
SIDDIQUE Fathma, SAKIB Shadman, and SIDDIQUE Md Abu Bakr, “Recognition of handwritten digit using convolutional neural network in python with tensorflow and comparison of performance for various hidden layers,” In: 2019 5th international conference on advances in electrical engineering (ICAEE), IEEE, 2019, pp. 541–546. https://doi.org/10.1109/ICAEE48663.2019.8975496
ANG Zelun and LIU Jyh-Charn, “Translating math formula images to LaTeX sequences using deep neural networks with sequence-level training,”. International Journal on Document Analysis and Recognition (IJDAR), 2021, 24.1–2: 63–75. https://doi.org/10.1007/s10032-020-00360-2
HEROUT Adam, et al., “Review of Hough transform for line detection,” Real-Time Detection of Lines and Grids: By PClines and Other Approaches, 2013, 3–16. https://doi.org/10.1007/978-1-4471-4414-4_2
ADEM Kemal, “Impact of activation functions and number of layers on detection of exudates using circular Hough transform and convolutional neural networks,” Expert Systems with Applications, 2022, 203: 117583. https://doi.org/10.1016/j.eswa.2022.117583
ZHAO Kai, et al., “Deep hough transform for semantic line detection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 44.9: 4793–4806.
Authors who publish with this journal agree to the following terms:
Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).