임펄스 노이즈 환경에서 변형된 Min-max 필터를 이용한 영상복원에 관한 연구

Abstract

Recently, thanks to the development of digital technology and information-communication technology, digital image device and related technology have been rapidly improved. Especially in digital image processing including image compression, image restoration, and image recognition, a lot of researches have progressed and noise removal skills of image restoration are essentially needed for getting a high quality image. These noise removal skills target the algorithm development to remove the variable noises efficiently which happen in the procedures of the data transmission. Generally, impulse noise and AWGN(additive white gaussian noise) are considered as representative noises which contaminate the digital image, and many considerable solutions are developed for the various types of noise caused by various factors. In the impulse noise environment, SM(standard median) filter is simple and it shows a great performance. Also, WM(weighed median) filter putting the weight in each pixel value and its modified CWM(center-weighted median) filter were proposed to enhance the image details. And various filters which applicate the SM filter and many modified median filters were proposed, so there has been a lot of progress in the impulse noise removal and the improvement of edge preserving performance. Although SM filter represents a good performance in the noise removal of image, it is insufficient at the edge preserving that is the other factor of image qualities. So, it can not be applied to the field that importantly considers the image details. Accordingly, modified median filter algorithms have been proposed to complement the defect, however they show the limit of increasing the image processing time a lot as repeatedly processing the computation. Therefore, it is needed the algorithm that removes noise in image and performs high speed computation while preserving the detail, and until now methods based on the min-max operation satisfied these conditions. According to these, in this paper, we enhanced the features of detail preservation of image and impulse noise removal by proposing a method based on modified min-max operation scheme. And we compared the efficiency of proposed filter with existing median filters and min-max based filter by simulation.