JBT Journal
Advanced

Raqamli Transformatsiya va Sun’iy Intellekt

DERMASKOPIK TASVIRLARDAN SOCHGA OID ARTEFAKTLARNI BARTARAF QILISHNING TRANSFORMER TEXNOLOGIYALARI VA TASVIR QISMLARINI TIKLASH(INPAINTING) GA ASOSLANGAN ALGORITMI

Akbaraliyev Baxtiyorjon Bakirovich1, Zaripov Fazilbek2, Gulmirzayeva Go‘zal Alisher qizi2

1 Andijon davlat universiteti

2 Toshkent axborot texnologiyalari universiteti

Zenodo DOI: 10.5281/zenodo.20623942

Keywords

SegFormer Dermaskopik tasvir chuqur o‘qitish soch artefaktlari tasvir qismlarini tiklash

Abstract

Dermaskopik tasvirlar teri o‘smalarini erta aniqlash va differensial tashxislashda keng qo‘llaniladi. Biroq, tasvirlarda uchraydigan soch artefaktlari diagnostika jarayonida jiddiy muammolarga olib keladi. Soch tolalari pigment tarmoqlari, shikastlanish chegaralari va rang gradiyentlarini qoplashi natijasida ham sun’iy intellekt (SI)ga asoslangan modellar, ham shifokorlarning vizual baholashga salbiy ta’sir qiladi. Ushbu tadqiqotda dermaskopik tasvirlardan soch artefaktlarni olib tashlash uchun ikki bosqichli chuqur o‘qitishga asoslangan yondashuv taklif qilindi. Birinchi bosqichda soch tolasi va uning soyalari SegFormerWithDropout arxitekturasi yordamida segmentatsiya qilinadi. Bunda dropout mexanizmi modellarning barqarorligini oshiradi, shadow-aware kengaytirilgan kanal esa soyalarni niqoblash imkonini beradi. Ikkinchi bosqichda gibrid Mask-Aware Transformer(MAT)+(Large Mask Inpainting with Fourier Convolutions) LaMa tasvir qismlarini tiklash generatori global va lokal kontekst asosida niqoblangan joylarni tiklash, artefaktlarsiz klinik tuzilmalarni (pigment tarmog‘i, rang o‘zgarishlari, chegaralar) qayta tiklash uchun qo‘llaniladi. Taklif qilingan yondashuv HAM10000 ma’lumotlar to‘plamida sinovdan o‘tkazildi va Dice (0.967), IoU (0.935) va SSIM (0.915) kabi yuqori ko‘rsatkichlariga erishildi. Olingan natijalar shuni ko‘rsatdiki, taklif qilingan usul dermaskopik tasvirlar sifatini oshirish orqali diagnostika tizimlarining umumiy samaradorligini oshirishi mumkin.

Views: 120 Published: 2026-06-10 Pages: 111-119 Volume: 4 Issue: 3

References

  1. K. Delibasis, K. Moutselos, E. Vorgiazidou, and I. Maglogiannis, ‘Automated hair removal in dermoscopy images using shallow and deep learning neural architectures’, Computer Methods and Programs in Biomedicine Update, vol. 4, p. 100109, 2023, doi: 10.1016/j.cmpbup.2023.100109.
  2. E. Xie, W. Wang, Z. Yu, A. Anandkumar, J. M. Alvarez, and P. Luo, ‘SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers’, 2021, arXiv. doi: 10.48550/ARXIV.2105.15203.
  3. L. Talavera-Martínez, P. Bibiloni, and M. González-Hidalgo, ‘Hair Segmentation and Removal in Dermoscopic Images Using Deep Learning’, IEEE Access, vol. 9, pp. 2694–2704, Dec. 2020, doi: 10.1109/access.2020.3047258.
  4. G. Randi et al., ‘The European Cancer Information System: exploring linkages between indoor radon concentrations and data on cancer burden’, radon, Mar. 2022, doi: 10.35815/radon.v3.7607.
  5. W. Li, Z. Lin, K. Zhou, L. Qi, Y. Wang, and J. Jia, ‘MAT: Mask-Aware Transformer for Large Hole Image Inpainting’, in 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA: IEEE, Jun. 2022, pp. 10748–10758. doi: 10.1109/CVPR52688.2022.01049.
  6. D. Bardou, H. Bouaziz, L. Lv, and T. Zhang, ‘Hair removal in dermoscopy images using variational autoencoders’, Skin Research and Technology, vol. 28, no. 3, pp. 445–454, May 2022, doi: 10.1111/srt.13145.
  7. W. Li, A. N. Joseph Raj, T. Tjahjadi, and Z. Zhuang, ‘Digital hair removal by deep learning for skin lesion segmentation’, Pattern Recognition, vol. 117, p. 107994, Sep. 2021, doi: 10.1016/j.patcog.2021.107994.
  8. J. K. Winkler et al., ‘Association between different scale bars in dermoscopic images and diagnostic performance of a market-approved deep learning convolutional neural network for melanoma recognition’, European Journal of Cancer, vol. 145, pp. 146–154, Mar. 2021, doi: 10.1016/j.ejca.2020.12.010.
  9. L. Jütte, H. Patel, and B. Roth, ‘Advancing dermoscopy through a synthetic hair benchmark dataset and deep learning-based hair removal’, Journal of Biomedical Optics, vol. 29, no. 11, pp. 116003–116003, Nov. 2024, doi: 10.1117/1.jbo.29.11.116003.
  10. J. A. A. Salido and C. Ruiz, ‘Using morphological operators and inpainting for hair removal in dermoscopic images’, in Proceedings of the Computer Graphics International Conference, Yokohama Japan: ACM, Jun. 2017, pp. 1–6. doi: 10.1145/3095140.3095142.
  11. R. Kasmi et al., ‘SharpRazor: Automatic removal of hair and ruler marks from dermoscopy images’, Skin Research and Technology, vol. 29, no. 4, Mar. 2023, doi: 10.1111/srt.13203.
  12. D. Liu, C. Lu, H. Sun, and S. Gao, ‘NA-segformer: A multi-level transformer model based on neighborhood attention for colonoscopic polyp segmentation’, Sci Rep, vol. 14, no. 1, p. 22527, Sep. 2024, doi: 10.1038/s41598-024-74123-y.
  13. Somnathe, Prathamesh A., and P. P. Gumaste., ‘A review of existing hair removal methods in dermoscopic images’, IOSR Journal of Electronics and Communication Engineering (IOSRJECE), vol. 1, pp. 73–76, 2015.
  14. S. Mocellin and C. R. Rossi, ‘The Melanoma Molecular Map Project’, Melanoma Research, vol. 18, no. 3, pp. 163–165, Jun. 2008, doi: 10.1097/CMR.0b013e328300c50b.
  15. R. Suvorov et al., ‘Resolution-robust Large Mask Inpainting with Fourier Convolutions’, in 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Waikoloa, HI, USA: IEEE, Jan. 2022, pp. 3172–3182. doi: 10.1109/WACV51458.2022.00323.
  16. Z. Wang et al., ‘Influence of hair presence on dermoscopic image analysis by AI in skin lesion diagnosis’, Computers in Biology and Medicine, vol. 183, p. 109335, Dec. 2024, doi: 10.1016/j.compbiomed.2024.109335.
  17. E. M. Onyema et al., ‘Deep learning model for hair artifact removal and Mpox skin lesion analysis and detection’, Sci Rep, vol. 15, no. 1, p. 21212, Jul. 2025, doi: 10.1038/s41598-025-05324-2.
  18. T. Sourget, S. N. Hasany, F. Mériaudeau, and C. Petitjean, ‘Can SegFormer be a True Competitor to U-Net for Medical Image Segmentation?’, in Medical Image Understanding and Analysis, vol. 14122, G. Waiter, T. Lambrou, G. Leontidis, N. Oren, T. Morris, and S. Gordon, Eds, in Lecture Notes in Computer Science, vol. 14122. , Cham: Springer Nature Switzerland, 2024, pp. 111–118. doi: 10.1007/978-3-031-48593-0_8.