Degraded image restoration using neural networks

Image processing systems are used in different applications. Blurred and noisy images formed for various reasons, many methods have been developed to reduce adverse effects. In order to determine the quality of the picture to compare the original and degraded state created a database of various statistical statements. This database contains images encoded taught artificial neural networks, artificial neural network simulations due to the degraded image is original or is within the database, or may decide.System proposed in this thesis, learned a lot of artificial neural network applied to different rates of three different kinds of interferences which are gaussian, blurring and speckle. It also learned the original state. After the neural network simulation examples taught, inside and outside of the database you selected, can be correctly predicted.Key words: Image Restoration, Blur, Image noise, Neural Network