A Study of the Effects of Gaussian Noise on Image Features

Abstract

In this paper, the effect of noise on the features of digital images has been tested. Since most of the computer and communication systems can be affected by Gaussian noise which may come from different natural sources, it was interesting to measure the effect of this noise on image features. For this purpose, a data set of several images was used with Gaussian noise to generate the noisy images. Different Mean and Variance values were used each time with the same data set to measure wide variety of noise parameters. Then, four different enhancement filters were used to remove the Gaussian noise. A total of 10 features were selected for this study. The features from original images, noisy images, and enhanced images are measured and compared. The results show that some filters offended the features of the image more than the noise itself, and all the filters have similar effects on Texture features, Entropy, and two of Wavelet-based features. According to the Texture Features, the Average enhancement filter came up with the best results, but according to the Wavelet-based Features, the Motion enhancement filter came up with the best results. The software used in this paper is Matlab 2013. ‫السالم‬ ‫عبد‬ ‫محمد‬ ‫أمين‬ 1 ‫فاضل‬ ‫فريدون‬ ‫أحمد‬ ، 2 2,1 ‫كمية‬ / ‫كركوك‬ ‫جامعة‬ ‫با‬ ‫الكهر‬ ‫قسم‬ / ‫الهندسة‬ ‫ء‬ ameen.selami@yahoo.com 1 amet83@yahoo.com 2 :‫البحث‬ ‫استالم‬ ‫تاريخ‬ 55 / 55 / 5155 :‫البحث‬ ‫قبول‬ ‫تاريخ‬ 55 / 4 / 5156 ‫الممخص‬ ‫تم‬ ،‫البحث‬ ‫ىذا‬ ‫في‬ ‫اء‬ ‫اجر‬ ‫تأثير‬ ‫اختبار‬ ‫ا‬ ‫أنظمة‬ ‫معظم‬ ‫أن‬ ‫حيث‬ ‫الرقمية‬ ‫الصور‬ ‫ات‬ ‫ميز‬ ‫عمى‬ ‫الضوضاء‬ ‫االتصاالت‬ ‫و‬ ‫لحاسبات‬ ‫معرضة‬ ‫لم‬ ‫ب‬ ‫تأثر‬ ‫ـ‬ ‫بالـ‬ ‫المعروفة‬ ‫الضوضاء‬ (Gaussian Noise ‫التي‬) ‫ف‬ ‫مختمفة‬ ‫طبيعية‬ ‫مصادر‬ ‫من‬ ‫تأتي‬ ‫قد‬ ‫المثير‬ ‫من‬ ‫كان‬ .‫الصور‬ ‫ات‬ ‫ميز‬ ‫عمى‬ ‫الضوضاء‬ ‫ىذه‬ ‫تأثير‬ ‫قياس‬ ‫يتم‬ ‫أن‬ ‫لالىتمام‬ ،‫الغرض‬ ‫ليذا‬ ‫من‬ ‫مجموعة‬ ‫استخدام‬ ‫تم‬ ‫ال‬ ‫صور‬ ‫المعرضة‬ ‫لـ‬ (Gaussian Noise ‫ا‬ ‫لتوليد‬) ‫لصور‬ ‫لم‬ ‫مختمفة‬ ‫قيم‬ ‫مع‬ ‫المشوىة‬ ‫و‬ ‫متوسط‬ ‫التباين‬ ‫الصور‬ ‫مجموعة‬ ‫نفس‬ ‫مع‬ ‫مرة‬ ‫كل‬ ‫في‬ ‫الجل‬ ‫وذلك‬ ‫قياس‬ ‫الضوضاء‬ ‫ىذه‬ ‫تاثير‬ ‫من‬ ‫متنوعة‬ ‫حاالت‬ ‫الضوضاء‬ ‫الة‬ ‫إلز‬ ‫مختمفة‬ ‫تعزيز‬ ‫مرشحات‬ ‫أربع‬ ‫استخدمت‬ ،‫ثم‬. (Gaussian Noise) ‫من‬ ‫مجموعة‬ ‫اختيرت‬ ‫و‬ 11 .‫اسة‬ ‫الدر‬ ‫ليذه‬ ‫ات‬ ‫ميز‬ ‫ان‬ ‫األ‬ ‫الصور‬ ‫من‬ ‫ات‬ ‫الميز‬ ‫الصور‬ ، ‫صمية‬ ‫المشوىة‬ ‫ال‬ ‫و‬ ، ‫بينيا‬ ‫المقارنة‬ ‫وتمت‬ ‫احتسبت‬ ‫قد‬ ‫المعززة‬ ‫صور‬ ‫ات‬ ‫ميز‬ ‫عمى‬ ‫اثرت‬ ‫التعزيز‬ ‫مرشحات‬ ‫بعض‬ ‫ان‬ ‫النتائج‬ ‫أظيرت‬. (‫الضوضاء‬ ‫من‬ ‫اكثر‬ ‫الصور‬ Gaussian Noise ‫أ‬ ‫إلى‬ ‫اضافة‬ ،‫نفسيا‬) ‫عمى‬ ‫نفسو‬ ‫التاثير‬ ‫ليا‬ ‫التعزيز‬ ‫مرشحات‬ ‫بعض‬ ‫ن‬ ‫ال‬ ‫ات‬ ‫ميز‬ ‫ـ‬ (Texture ‫ال‬ ‫و‬) ‫ـ‬ (Entropy ‫و‬) ‫من‬ ‫اثنان‬ ‫ال‬ (‫التحويل‬ ‫عمى‬ ‫المعتمدة‬ ‫ات‬ ‫ميز‬ Wavelet ‫استنادا‬ .) ‫إلى‬ ‫ال‬ ‫ات‬ ‫ميز‬ ‫ـ‬ (Texture ‫استنادا‬ ‫بينما‬ ،‫النتائج‬ ‫افضل‬ ‫اعطت‬ ‫المتوسط‬ ‫التعزيز‬ ‫مرشح‬ ،) ‫إلى‬ ‫ال‬ ‫التحويل‬ ‫عمى‬ ‫المعتمدة‬ ‫ات‬ ‫ميز‬ (Wavelet ‫الب‬ .‫النتائج‬ ‫افضل‬ ‫اعطت‬ ‫الحركة‬ ‫تعزيز‬ ‫مرشح‬ ‫فان‬) ‫ر‬ ‫الب‬ ‫ىذا‬ ‫في‬ ‫استخدامو‬ ‫تم‬ ‫الذي‬ ‫نامج‬ ‫ىو‬ ‫حث‬ Matlab 2013. ‫ال‬ ‫كممات‬ ‫ال‬ :‫دالة‬ ‫الضوضاء‬ , ‫الصور‬ ‫ات‬ ‫ميز‬ Gaussian Noise ‫يز‬ ‫التعز‬ ‫مرشحات‬ , , ‫التحويل‬ (Wavelet) .