Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning

Kazanç Murat; Ensari Tolga; Dağtekin Mustafa

doi:https://dx.doi.org/10.26650/acin.1258719

Araştırma Makalesi

DOI :10.26650/acin.1258719 IUP :10.26650/acin.1258719 Tam Metin (PDF)

Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning

Murat Kazanç, Tolga Ensari, Mustafa Dağtekin

A few decades ago, people used printed resources such as books and magazines to learn. With the development of technology, digital documents have replaced printed resources. These documents can occur in the form of images or various text formats. Many different applications exist for preparing digital documents, one of these being LaTeX. LaTeX is a document preparation system and typesetting software that is used especially in the field of scientific publications and mathematics for preparing high quality documents. When preparing a document using LaTeX, the content is made ready using a markup language, which creates difficulties for some users. However, one of the main advantages of using the LaTeX system is that it distinguishes the document’s content from its formatting. Once the content is created, the formatting can be easily replaced. Generating LaTeX code from an image-formatted document requires both the use of computer vision and NLP. This study discovers the boundaries (blocks) of the places where text, tables, and figures are located on an image before making a text classification using the natural language processing methods of these blocks. The next stage of the study determines the reading order to enable meaningful flow. The final stage of the study produces a LaTeX code using the obtained information.

Anahtar Kelimeler: Computer vision, text classification, reading order, machine learning

DOI :10.26650/acin.1258719 IUP :10.26650/acin.1258719 Tam Metin (PDF)

Resim Formatındaki Dokümanların Bilgisayarlı Görü, Doğal Dil İşleme ve Makine Öğrenmesi Kullanılarak Latex Formatına Dönüştürülmesi

Murat Kazanç, Tolga Ensari, Mustafa Dağtekin

Birkaç on yıl önce insanlar bilgi edinmek için kitap ve dergi gibi basılı kaynakları kullanmaktaydılar. Teknolojinin gelişmesi ile basılı kaynakların yerini dĳital dokümanlar almıştır. Bu dokümanlar görüntü biçiminde veya farklı metin formatları şeklinde olabilmektedir. Dĳital dokümanları hazırlamak için birçok farklı uygulama bulunmaktadır. Bunlardan bir tanesi LaTex’ tir. LaTex doküman hazırlama sistemi ve dizgi yazılımıdır. Yüksek kalitede dokümanlar hazırlamak için özellikle bilimsel yayınlar ve matematik alanında kullanılmaktadır. LaTex ile doküman hazırlanırken içerik bir işaretleme dili kullanılarak hazırlanılmaktadır. Bu durum bazı kullanıcılar için bir zorluk oluşturmaktadır. Ancak LaTex sistemini kullanmanın ana avantajlarından biri doküman içeriğini biçimlendirmeden ayırmasıdır. Bir kere içerik oluşturulduktan sonra biçimlendirme kolaylıkla değiştirilebilmektedir. Görüntü formatındaki bir dokümandan LaTex kodunun üretilmesi bilgisayarlı görü ve doğal dil işleme alanlarının birlikte kullanılmasını gerektirmektedir. Bu çalışmada öncelikle görüntü üzerinde metin, tablo ve şekillerin bulunduğu yerlerin sınırları (bloklar) tespit edilmiştir. Sonrasında bulunan bu blokların doğal dil işleme metotları kullanılarak metin sınıflama yapılmıştır. Bir sonraki aşamada anlam akışının bozulmaması için okuma sırası tespit edilmiştir. Son aşamada elde edilen bilgiler kullanılarak LaTex kodu üretilmiştir.

Anahtar Kelimeler: Bilgisayarlı görü, metin sınıflama, okuma sırası, makine öğrenmesi

Referanslar

Akpan, U. I., & Starkey, A. (2021). Review of classification algorithms with changing inter-class distances. Machine Learning with Applications, 4, 100031. https://doi.org/10.1016/j.mlwa.2021.100031 google scholar
Ali, F., Kwak, K.-S., & Kim, Y.-G. (2016). Opinion mining based on fuzzy domain ontology and Support Vector Machine: A proposal to automate online review classification. Applied Soft Computing, 47, 235-250. https://doi.Org/10.1016/j.asoc.2016.06.003 google scholar
Clark, C., & Divvala, S. (2016). PDFFigures 2.0. Proceedings of the 16th ACM/IEEE-CS on Joint Conference on Digital Libraries, 143-152. https://doi.org/10.1145/2910896.2910904 google scholar
CTAN Team. (n.d.). What are TEX and its friends? Retrieved May 8, 2022, from https://www.ctan.org/tex google scholar
Deivalakshmi, S., Palanisamy, P., & Vishwanathan, G. (2013). A novel method for text and non-text segmentation in document images. 2013 International Conference on Communication and Signal Processing, 255-259. https://doi.org/10.1109/iccsp.2013.6577054 google scholar
Deng, Y., Rosenberg, D., & Mann, G. (2019). Challenges in End-to-End Neural Scientific Table Recognition. 2019 International Conference on Document Analysis and Recognition (ICDAR), 894-901. https://doi.org/10.1109/ICDAR.2019.00148 google scholar
Ding, H., Chen, K., & Huo, Q. (2019). Compressing CNN-DBLSTM models for OCR with teacher-student learning and Tucker decomposition. Pattern Recognition, 96, 106957. https://doi.org/10.1016/j.patcog.2019.07.002 google scholar
Doğan, M. İ., Orman, A., Örkcü, M., & Örkcü, H. H. (2019). Çok gruplu sınıflandırma problemlerine regresyon analizi ve matematiksel programlama tabanlı yeni bir yaklaşım. Gazi Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi. https://doi.org/10.17341/gazimmfd.571643 google scholar
Kavasidis, I., Pino, C., Palazzo, S., Rundo, F., Giordano, D., Messina, P., & Spampinato, C. (2019). A Saliency-Based Convolutional Neural Network for Table and Chart Detection in Digitized Documents. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 11752 LNCS, 292-302. https://doi.org/10.1007/978-3-030-30645-8_27 google scholar
Klatsky, S. (2003). WYSIWYG. Aesthetic Surgery Journal, 23(4), 274-275. https://doi.org/10.1016/S1090-820X(03)00150-X google scholar
Li, M., Cui, L., Huang, S., Wei, F., Zhou, M., & Li, Z. (2019). TableBank: A Benchmark Dataset for Table Detection and Recognition. http://arxiv.org/abs/1903.01949 google scholar
Navada, A., Ansari, A. N., Patil, S., & Sonkamble, B. A. (2011). Overview of use of decision tree algorithms in machine learning. 2011 IEEE Control and System Graduate Research Colloquium, 37-42. https://doi.org/10.1109/ICSGRC.2011.5991826 google scholar
Otsu, N. (1979). A Threshold Selection Method from Gray-Level Histograms. IEEE Transactions on Systems, Man, and Cybernetics, 9(1), 62-66. https://doi.org/10.1109/TSMC.1979.4310076 google scholar
Pang, N., Yang, C., Zhu, X., Li, J., & Yin, X.-C. (2021). Global Context-Based Network with Transformer for Image2latex. 2020 25th International Conference on Pattern Recognition (ICPR), 4650-4656. https://doi.org/10.1109/ICPR48806.2021.9412072 google scholar
PRImA. (n.d.). Retrieved May 22, 2022, from https://www.primaresearch.org/ google scholar
Recommendation ITU-R BT.601-7. (2011, March). https://www.itu.int/dmspubrec/itu-r/rec/bt/R-REC-BT.601-7-201103-I!!PDF-E.pdf google scholar
Safnuk, B., & Hu, G. (2018). Reconstructing LaTeX Source Files from Generated PDFs - a Neural Network Approach. 2018 IEEE 16th International Conference on Industrial Informatics (INDIN), 890-895. https://doi.org/10.1109/INDIN.2018.8472050 google scholar
Shen, Z., Zhang, R., Dell, M., Lee, B. C. G., Carlson, J., & Li, W. (2021). LayoutParser: A Unified Toolkit for Deep Learning Based Document Image Analysis. http://arxiv.org/abs/2103.15348 google scholar
Uğuz, H. (2011). A two-stage feature selection method for text categorization by using information gain, principal component analysis and genetic algorithm. Knowledge-Based Systems, 24(7), 1024-1032. https://doi.org/10.1016/j.knosys.2011.04.014 google scholar
Wang, Z., & Liu, J. C. (2021). Translating math formula images to LaTeX sequences using deep neural networks with sequence-level training. International Journal on Document Analysis and Recognition, 24(1-2), 63-75. https://doi.org/10.1007/s10032-020-00360-2 google scholar
Wang, Z., Xu, Y., Cui, L., Shang, J., & Wei, F. (2021). LayoutReader: Pre-training of Text and Layout for Reading Order Detection. http://arxiv.org/abs/2108.11591 google scholar
Wang, Z., Yang, J., Jin, H., Shechtman, E., Agarwala, A., Brandt, J., & Huang, T. S. (2015). DeepFont: Identify Your Font from An Image. Proceedings of the 23rd ACM International Conference on Multimedia, 451-459. https://doi.org/10.1145/2733373 google scholar
Xu, C., Shi, C., Bi, H., Liu, C., Yuan, Y., Guo, H., & Chen, Y. (2021). A Page Object Detection Method Based on Mask R-CNN. IEEE Access, 9, 143448-143457. https://doi.org/10.1109/ACCESS.2021.3121152 google scholar
Xu, Y., Li, M., Cui, L., Huang, S., Wei, F., & Zhou, M. (2020). LayoutLM: Pre-training of Text and Layout for Document Image Under-standing. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 20, 1192-1200. https://doi.org/10.1145/3394486.3403172 google scholar
Zhong, X., Tang, J., & Yepes, A. J. (2019). PubLayNet: largest dataset ever for document layout analysis. http://arxiv.org/abs/1908.07836 google scholar

Atıflar

Biçimlendirilmiş bir atıfı kopyalayıp yapıştırın veya seçtiğiniz biçimde dışa aktarmak için seçeneklerden birini kullanın

DIŞA AKTAR

APA

Kazanç, M., Ensari, T., & Dağtekin, M. (2023). Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning. Acta Infologica, 7(2), 253-266. https://doi.org/10.26650/acin.1258719

AMA

Kazanç M, Ensari T, Dağtekin M. Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning. Acta Infologica. 2023;7(2):253-266. https://doi.org/10.26650/acin.1258719

ABNT

Kazanç, M.; Ensari, T.; Dağtekin, M. Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning. Acta Infologica, [Publisher Location], v. 7, n. 2, p. 253-266, 2023.

Chicago: Author-Date Style

Kazanç, Murat, and Tolga Ensari and Mustafa Dağtekin. 2023. “Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning.” Acta Infologica 7, no. 2: 253-266. https://doi.org/10.26650/acin.1258719

Chicago: Humanities Style

Kazanç, Murat, and Tolga Ensari and Mustafa Dağtekin. “Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning.” Acta Infologica 7, no. 2 (May. 2024): 253-266. https://doi.org/10.26650/acin.1258719

Harvard: Australian Style

Kazanç, M & Ensari, T & Dağtekin, M 2023, 'Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning', Acta Infologica, vol. 7, no. 2, pp. 253-266, viewed 3 May. 2024, https://doi.org/10.26650/acin.1258719

Harvard: Author-Date Style

Kazanç, M. and Ensari, T. and Dağtekin, M. (2023) ‘Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning’, Acta Infologica, 7(2), pp. 253-266. https://doi.org/10.26650/acin.1258719 (3 May. 2024).

MLA

Kazanç, Murat, and Tolga Ensari and Mustafa Dağtekin. “Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning.” Acta Infologica, vol. 7, no. 2, 2023, pp. 253-266. [Database Container], https://doi.org/10.26650/acin.1258719

Vancouver

Kazanç M, Ensari T, Dağtekin M. Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning. Acta Infologica [Internet]. 3 May. 2024 [cited 3 May. 2024];7(2):253-266. Available from: https://doi.org/10.26650/acin.1258719 doi: 10.26650/acin.1258719

ISNAD

Kazanç, Murat - Ensari, Tolga - Dağtekin, Mustafa. “Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning”. Acta Infologica 7/2 (May. 2024): 253-266. https://doi.org/10.26650/acin.1258719

Cilt 7, Sayı 22023, S. 253-266

ZAMAN ÇİZELGESİ

Gönderim	01.03.2023
Kabul	19.09.2023
Çevrimiçi Yayınlanma	26.10.2023

LİSANS

Attribution-NonCommercial (CC BY-NC)

This license lets others remix, tweak, and build upon your work non-commercially, and although their new works must also acknowledge you and be non-commercial, they don’t have to license their derivative works on the same terms.

Acta Infologica

Araştırma Makalesi

Converting Image Files to LaTeX Format Using Computer Vision, Natural Language Processing, and Machine Learning

Resim Formatındaki Dokümanların Bilgisayarlı Görü, Doğal Dil İşleme ve Makine Öğrenmesi Kullanılarak Latex Formatına Dönüştürülmesi

PDF Görünüm

Referanslar

Atıflar

DIŞA AKTAR

APA

AMA

ABNT

Chicago: Author-Date Style

Chicago: Humanities Style

Harvard: Australian Style

Harvard: Author-Date Style

MLA

Vancouver

ISNAD

ZAMAN ÇİZELGESİ

LİSANS

PAYLAŞ