پیوندهای مفید
آمار
| تعداد نشریات | 33 |
| تعداد شماره‌ها | 886 |
| تعداد مقالات | 8,690 |
| تعداد مشاهده مقاله | 55,111,849 |
| تعداد دریافت فایل اصل مقاله | 10,383,942 |
A smoothing-based conjugate gradient approach for image restoration | ||
| Journal of Mathematical Modeling | ||
| مقالات آماده انتشار، اصلاح شده برای چاپ، انتشار آنلاین از تاریخ 06 آبان 1405 اصل مقاله (1.16 M) | ||
| نوع مقاله: Research Article | ||
| شناسه دیجیتال (DOI): 10.22124/jmm.2026.35108.3280 | ||
| نویسندگان | ||
| Nurullah Yilmaz* 1؛ Ahmet Sahiner1؛ Gulden Kapusuz Gokdogan2 | ||
| 1Department of Mathematics, Suleyman Demirel University, Isparta, Türkiye | ||
| 2Department of Basic Sciences, Isparta University of Applied Sciences, Isparta, Türkiye | ||
| چکیده | ||
| This paper studies image restoration using Total Variation, non-convex Lipschitz, and non-Lipschitz regularizers. We use indicator-based smoothing functions (local polynomial and global arctangent) within a Smoothing Conjugate Gradient Algorithm (SCGA) and an inner-outer continuation framework. The search direction satisfies the sufficient descent condition. For locally Lipschitz models, global convergence to Clarke stationary points is established. For the non-Lipschitz model, the limit points are shown to satisfy asymptotic $\varepsilon$-stationarity. Numerical experiments over independent noise realizations evaluate the proposed variants against baseline solvers. The results indicate that the non-convex Lipschitz and non-Lipschitz models obtain higher PSNR under heavy noise, whereas baseline methods maintain competitive SSIM and CPU times in impulse noise settings. | ||
| کلیدواژه‌ها | ||
| Smoothing techniques؛ conjugate gradient method؛ non-smooth optimization؛ image restoration | ||
|
آمار تعداد مشاهده مقاله: 7 تعداد دریافت فایل اصل مقاله: 6 |
||