Raqamli Transformatsiya va Sun’iy Intellekt
WAVELET-TRANSFORMATSIYA HH QUYI-TASMASINING KOEFFITSIENTLAR STATISTIKASI ASOSIDA TASVIRDAGI SHOVQIN DARAJASINI ETALONSIZ ANIQLASH MEZONI
1 Muhammad al-Xorazmiy nomidagi TATU
Zenodo DOI: 10.5281/zenodo.23885056
Keywords
Abstract
Maqolada raqamli tasvirdagi additiv shovqin darajasini etalon tasvirsiz baholash masalasi ko’rib chiqiladi. Mavjud shovqinni bartaraf etish usullarining aksariyati (BM3D, wavelet-chegaralash, FFDNet kabi) samarali ishlashi uchun shovqin darajasi parametri σ ni oldindan bilishni talab qiladi, biroq real sharoitlarda etalon tasvir mavjud emas. Ushbu ishda vaqt-chastota (wavelet) tahliliga asoslangan, bir bosqichli Haar diskret wavelet transformatsiyasining yuqori-chastotali HH quyi-tasmasi koeffitsientlari medianasiga tayanuvchi klassik Donoho-Johnstone MAD (Median Absolute Deviation - medianadan mutlaq og'ish) baholovchisi qo’llaniladi va uning amaliy aniqligi standart test-tasvirlarda tajribaviy tekshiriladi. Sakkizta xilma-xil standart test-tasvirda (Cameraman, Coins, Moon, Page, Checkerboard, Astronaut, Brick, Grass), beshta shovqin darajasida (σ = 10, 15, 25, 30, 50), har birida 20 mustaqil realizatsiya bo’yicha o’tkazilgan tajriba shuni ko’rsatadiki, baholash xatoligi shovqin darajasi oshgani sayin monoton kamayadi (σ = 10 da o’rtacha 15,3%, σ = 50 da 1,3%) va tasvir tarkibidagi tekstura zichligiga kuchli bog’liq (toza tasvirning bazaviy HH-MAD darajasi bilan xatolik o’rtasida Pirson korrelyatsiyasi r = 0,97). Aniqlangan bog’liqlik mezonning ishonchlilik chegaralarini belgilaydi va uni tasvir turini aniqlovchi qo’shimcha mezonlar bilan boyitish zarurligini asoslaydi.
References
- Pimpalkhute V.A., Page R., Kothari A., Bhurchandi K.M., Kamble V.M. Digital Image Noise Estimation Using DWT Coefficients // IEEE Transactions on Image Processing. — 2021. — Vol. 30. — P. 1962–1972.
- Muhamediyeva D.T., Mamatov N.S., Ametova G. Using quantum computing for image processing and inversion // AIP Conference Proceedings. — 2025. — Vol. 3377, Art. 060010.
- Mittal A., Moorthy A.K., Bovik A.C. No-Reference Image Quality Assessment in the Spatial Domain (BRISQUE) // IEEE Transactions on Image Processing. — 2012.
- Bharati S., Khan T.Z., Podder P. A Comparative Analysis of Image Denoising Problem: Noise Models, Denoising Filters and Applications // Cognitive Internet of Medical Things for Smart Healthcare (Studies in Systems, Decision and Control, vol. 311). — Springer, Cham, 2021. — P. 49–66.
- Mamatov N.S., Erejepov K.K., Narzullayev I.S., Jalelova M.M. Yuz tasvirlarini segmentlashning an'anaviy va mashinali o'qitish usullari // Raqamli texnologiyalarning nazariy va amaliy masalalari xalqaro jurnali. — 2024. — T. 7, №1. — B. 24–30.
- Liu P., Zhang H., Zhang K., Lin L., Zuo W. Multi-level Wavelet-CNN for Image Restoration // arXiv:1805.07071. — 2018.
- Narzullayev H. Tasvirlarga raqamli ishlov berish // Talqin va tadqiqotlar. — 2023. — T. 1, №25.
- Dabov K., Foi A., Katkovnik V., Egiazarian K. Image Denoising by Sparse 3-D Transform-Domain Collaborative Filtering // IEEE Transactions on Image Processing. — 2007.
- Zhang K., Zuo W., Zhang L. FFDNet: Toward a Fast and Flexible Solution for CNN-Based Image Denoising // IEEE Transactions on Image Processing. — 2018.
- Donoho D.L., Johnstone I.M. Ideal spatial adaptation by wavelet shrinkage // Biometrika. — 1994.
- Norboyeva N. Tasvirlarni intellektuallashtirishda filtrlash jarayoni va baholash algoritmlari // Green Economy and Development. — 2024. — T. 2, №11.