EAST ALGORITMI ASOSIDA MURAKKAB FONLI TASVIRLARDAGI MATN HUDUDLARINI ANIQLASH MODELI
DOI:
https://doi.org/10.54613/ku.v18iC.1738Keywords:
EAST algoritmi, STR, scene text recognition, scene text detection, murakkab fonli tasvirlar, matn lokalizatsiyasi, matnni tanib olish, IoU, F1-score.Abstract
Ushbu ilmiy ishda murakkab fonli tasvirlardagi matn hududlarini aniqlash EAST algoritmi asosida amalga oshirildi. Turli xil tasvirlarda matnlar noteks joylashuvi va tasvirning sifati yorug‘lik yetishmasligi, shriftlarnining turli xil uslubda bo‘lishi kabi omillar ta’siri natijasida tasvirdagi matnlarni aniqlash qiyinchiliklari kelib chiqadi. Shunday sabablarga ko‘ra matni tanib olishda avval hududlari tanib olish va STR modelidan foydalanilish kerak bo‘ladi. Ushbu tadqiqotdan East algoritmi asosida hududlarni aniqlash va STR usuli bilan matnning mazmunini tanib olish modeli ishlab chiqilgan. Model quyidagi bosqichlardan iborat: birinchi bosqichda tasvirni qayta ishlash, EAST algortimiga uzatish, matn joylashgan joylashuvini aniqlash, keyingi bosqichda takroriy freymlarni aniqlash va uni filtrlash, bundan keyingi bosqichda esa ajratib olingan hududlarni alohida ajratib olish va STR orqali matnni tanib olishdan iborat. Modelning samaradorligini esa precision, recall, f1-score, IoU kabi aniqlash mezonlari yordamida baholash amalga oshirildi. Bunda EAST algoritmi tasvir ichidagi matn joylashgan hududlarni tezkor va aniq aniqlash imkonini beradi. STR modeli esa ajratib olingan matn hududlaridagi belgilar ketma-ketligini mazmunli matn ko‘rinishida tanib olishga xizmat qiladi. Taklif etilgan yondashuv murakkab fon, turli ranglar, notekis yoritilish va shrift xilma-xilligi mavjud bo‘lgan tasvirlarda ham samarali ishlashga yo‘naltirilgan. Tadqiqot jarayonida matn hududlarini aniqlash va tanib olish bosqichlarining o‘zaro bog‘liqligi model natijalariga bevosita ta’sir qilishi aniqlandi. Shuningdek, ajratilgan hududlarning sifati STR orqali tanib olish aniqligini oshirishda muhim omil hisoblanadi. Natijada, ishlab chiqilgan model tasvirlardagi matnlarni avtomatik aniqlash va ularning mazmunini tanib olish jarayonini takomillashtirishga xizmat qiladi.
Foydalanilgan adabiyotlar:
1. Long, S., He, X., & Yao, C. (2021). Scene text detection and recognition: The deep learning era. International Journal of Computer Vision, 129(1), 161–184.
2. Yao, C., Wu, J., Zhou, X., Zhang, C., Zhou, S., Cao, Z., & Yin, Q. (2015). Incidental scene text understanding: Recent progresses on ICDAR 2015 Robust Reading Competition Challenge 4. arXiv preprint arXiv:1511.09207.
3. Shi, B., Bai, X., & Yao, C. (2017). An end-to-end trainable neural network for image-based sequence recognition and its application to scene text recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(11), 2298–2304.
4. Zhou, X., Yao, C., Wen, H., Wang, Y., Zhou, S., He, W., & Liang, J. (2017). EAST: An efficient and accurate scene text detector. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 5551–5560.
5. Shi, B., Wang, X., Lyu, P., Yao, C., & Bai, X. (2016). Robust scene text recognition with automatic rectification. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 4168–4176.
6. Shi, B., Yang, M., Wang, X., Lyu, P., Yao, C., & Bai, X. (2019). ASTER: An attentional scene text recognizer with flexible rectification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 41(9), 2035–2048.
7. Busta, M., Neumann, L., & Matas, J. (2017). Deep TextSpotter: An end-to-end trainable scene text localization and recognition framework. Proceedings of the IEEE International Conference on Computer Vision, 2204–2212.
8. Liao, M., Wan, Z., Yao, C., Chen, K., & Bai, X. (2020). Real-time scene text detection with differentiable binarization. Proceedings of the AAAI Conference on Artificial Intelligence, 34(07), 11474–11481.
9. Wang, W., Xie, E., Li, X., Hou, W., Lu, T., Yu, G., & Shao, S. (2019). Shape robust text detection with progressive scale expansion network. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 9336–9345.
10. Bautista, D., & Atienza, R. (2022). Scene text recognition with permuted autoregressive sequence models. European Conference on Computer Vision, 178–196.
Downloads
Published
Iqtiboslik olish
Issue
Section
License
Copyright (c) 2026 QO‘QON UNIVERSITETI XABARNOMASI
This work is licensed under a Creative Commons Attribution 4.0 International License.