Optimalisasi Teknik Reduksi Noise: Studi Perbandingan Metode Filtering untuk Peningkatan Citra

Abstract

Digital image restoration is a critical aspect of image processing, as noise introduced during acquisition, transmission, or compression can degrade visual quality and reduce the accuracy of image information. The main challenge in noise reduction lies in suppressing disturbances without damaging important image details and structural features. This study aims to evaluate the effectiveness of Gaussian Filter, Median Filter, and Mean Filter, both individually and in combination, for noise reduction in digital images. The dataset consists of JPG images with a resolution of 4032×3024 pixels (12 MP), acquired using a smartphone camera and artificially contaminated with noise to simulate real-world conditions. Performance evaluation was conducted using noise standard deviation, Peak Signal-to-Noise Ratio (PSNR), and Structural Similarity Index Measure (SSIM). Experimental results indicate that the combination of the Median Filter and Gaussian Filter achieves the best overall performance, with a noise standard deviation of 88.08, a PSNR of 13.32 dB, and an SSIM of 0.15, demonstrating an optimal balance between noise reduction and structural preservation. The findings confirm that combined filtering approaches are more effective than single filters. Future research is recommended to explore advanced filtering methods such as Bilateral Filter, Wiener Filter, and adaptive filtering techniques under various noise conditions