Published October 3, 2026 | Version v1

PAST ANIQLIKDAGI QO'LYOZMA HUJJATLAR TASVIRLARINI TIKLASHGA MOSLASHTIRILGAN Real-ESRGAN MODELI

  • 1. Tayanch doktorant, «TIQXMMI» MTU huzuridagi FATI
  • 2. Texnika fanlari bo'yicha falsafa doktori (PhD), Bosh ilmiy xodim, O'zbekiston Respublikasi Prezidenti Administratsiyasi huzuridagi Iqtisodiy tadqiqotlar va islohotlar markazi
  • 3. Texnika fanlari bo'yicha falsafa doktori (PhD), Dotsent, Toshkent menejment va iqtisodiyot institute Muhandislik va axborot texnologiyalari kefadrasi
  • 4. Texnika fanlari doktori, Laboratoriya mudiri, «TIQXMMI» MTU huzuridagi FATI

Description

Ushbu maqolada universal generativ Real-ESRGAN modelini hujjat domeniga moslashtirish masalasi tadqiq etilgan. Dastlabki model uch yo'nalishda o'zgartirildi: degradatsiya modelining parametrlari qayta sozlandi, generator sig'imi ikki baravar kamaytirildi (8.79 mln parametr), qo'lyozma matnni tanib olish masalasining o'ziga xosligini hisobga oluvchi yo'qotish funksiyalari kiritildi. Model uch tildagi qo'lyozma matn tasvirlarining 208 202 ta bo'lagidan iborat korpusda o'qitildi. Dastlabki tasvirning eng murakkab degradatsiya darajasida taklif etilayotgan modifikatsiyadan foydalanilganda belgilarni tanib olishdagi xatoliklar ulushi 12.64 % dan 10.61 % gacha kamaydi. Taklif etilgan modelning qo'lyozma matnni tanib olish natijalariga ta'sir darajasini eng keng tarqalgan o'n ikkita usul bilan qiyosiy tahlil qilish uning ham tezkorlik (eng yaqin raqibidan 17 baravar tez), ham tanib olish aniqligi (birinchi o'rin) jihatidan yuqori samaradorligini ko'rsatdi.

Files

46_1300-330-340-Allayorov.pdf

Files (531.6 kB)

Name Size
md5:137ba60142c2714945404b6e1127cb3f
531.6 kB Preview Download

Additional details

References

  • Wang X., Xie L., Dong C., Shan Y. Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data // Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops. – 2021. – P. 1905–1914.
  • Dong C., Loy C.C., He K., Tang X. Image Super-Resolution Using Deep Convolutional Networks // IEEE Transactions on Pattern Analysis and Machine Intelligence. – 2016. – Vol. 38, No. 2. – P. 295–307.
  • Lim B., Son S., Kim H., Nah S., Lee K.M. Enhanced Deep Residual Networks for Single Image Super-Resolution // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. – 2017. – P. 1132–1140.
  • Ledig C., Theis L., Huszár F. et al. Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. – 2017. – P. 105–114.
  • Goodfellow I., Pouget-Abadie J., Mirza M. et al. Generative Adversarial Nets // Advances in Neural Information Processing Systems. – 2014. – Vol. 27. – P. 2672–2680.
  • Wang X., Yu K., Wu S. et al. ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks // Proceedings of the European Conference on Computer Vision Workshops. – 2018. – P. 63–79.
  • Zhang K., Liang J., Van Gool L., Timofte R. Designing a Practical Degradation Model for Deep Blind Image Super-Resolution // Proceedings of the IEEE/CVF International Conference on Computer Vision. – 2021. – P. 4771–4780.
  • Liang J., Cao J., Sun G. et al. SwinIR: Image Restoration Using Swin Transformer // Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops. – 2021. – P. 1833–1844.
  • Wang W., Xie E., Liu X. et al. Scene Text Image Super-Resolution in the Wild // Proceedings of the European Conference on Computer Vision. – 2020. – P. 650–666.
  • Shi B., Bai X., Yao C. 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. – 2017. – Vol. 39, No. 11. – P. 2298–2304.
  • Graves A., Fernández S., Gomez F., Schmidhuber J. Connectionist Temporal Classification: Labelling Unsegmented Sequence Data with Recurrent Neural Networks // Proceedings of the 23rd International Conference on Machine Learning. – 2006. – P. 369–376.
  • Blau Y., Michaeli T. The Perception-Distortion Tradeoff // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. – 2018. – P. 6228–6237.
  • Wang Z., Bovik A.C., Sheikh H.R., Simoncelli E.P. Image Quality Assessment: From Error Visibility to Structural Similarity // IEEE Transactions on Image Processing. – 2004. – Vol. 13, No. 4. – P. 600–612.
  • Simonyan K., Zisserman A. Very Deep Convolutional Networks for Large-Scale Image Recognition // International Conference on Learning Representations. – 2015. – 14 p.
  • Miyato T., Kataoka T., Koyama M., Yoshida Y. Spectral Normalization for Generative Adversarial Networks // International Conference on Learning Representations. – 2018. – 26 p.
  • Ronneberger O., Fischer P., Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation // Medical Image Computing and Computer-Assisted Intervention. – 2015. – P. 234–241.
  • Kingma D.P., Ba J. Adam: A Method for Stochastic Optimization // International Conference on Learning Representations. – 2015. – 15 p.
  • Otsu N. A Threshold Selection Method from Gray-Level Histograms // IEEE Transactions on Systems, Man, and Cybernetics. – 1979. – Vol. 9, No. 1. – P. 62–66.
  • Levenshtein V.I. Binary Codes Capable of Correcting Deletions, Insertions and Reversals // Soviet Physics Doklady. – 1966. – Vol. 10, No. 8. – P. 707–710.
  • Potyashin I., Kaprielova M., Chekhovich Yu. et al. HWR200: New Open Access Dataset of Handwritten Texts Images in Russian // Computational Linguistics and Intellectual Technologies: Papers from the Annual International Conference "Dialogue". – Moscow, 2023. – Iss. 22. – P. 452–458.
  • Sánchez J.A., Romero V., Toselli A.H., Vidal E. ICFHR2014 Competition on Handwritten Text Recognition on Transcriptorium Datasets // Proceedings of the 14th International Conference on Frontiers in Handwriting Recognition. – 2014. – P. 785–790.
  • Efron B., Tibshirani R.J. An Introduction to the Bootstrap. – New York: Chapman and Hall/CRC, 1994. – 456 p.