PENERAPAN METODE MID POINT FILTER MENGURANGI EFEK NOISE GAUSSIAN PADA CITRA BERBASIS ANDROID

Abstract

Noise in digital images, particularly Gaussian Noise, often poses a major challenge in maintaining the quality of visual information as it produces random speckles that can interfere with object identification and recognition processes. This problem becomes more complex on mobile devices, which are limited in computational power and storage capacity. This study proposes the application of the Midpoint Filter Method as a solution to reduce the effects of Gaussian Noise in Android-based images. The Midpoint Filter works by replacing each pixel value with the average of the maximum and minimum values within a filter window. This approach is chosen for its lightweight, efficient characteristics and its ability to reduce random noise without degrading important image details. The method is implemented through an Android application designed to load images, process them using the Midpoint Filter, and display a real-time comparison between the original and filtered images. Experimental results show that the proposed method significantly reduces Gaussian noise levels. In an experiment using an image with a resolution of 576×345 pixels and a file size of approximately 1 MB, the initial noise value of 2791 was successfully reduced to 1246 after applying the Midpoint Filter. Thus, the total noise reduction reached 1545, representing a 55.36% decrease, while maintaining the image resolution.