Open AccessDesigning Image Correction Software with the Wavelet Method
Author(s) -
Akbar Iskandar,
Dahlan Abdullah,
Mansyur,
Frédéric Herman
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1364/1/012028
Subject(s) - computer vision , filter (signal processing) , artificial intelligence , wavelet , computer science , noise (video) , transformation (genetics) , process (computing) , image (mathematics) , image processing , software , pixel , image restoration , biochemistry , chemistry , gene , programming language , operating system
Images can be produced by tools such as cameras, satellites or scanners. The process of transferring images from these devices can cause interference in the form of noise. Damage to photos or images is something that cannot be avoided, where this damage is usually caused by a time factor, a factor in someone’s accidental handling of a photo, or other external factors. This will cause the photo to be cured until information loss or artistic value occurs. Image processing is a solution to handle or manipulate an image, in other words in image processing is also offered a technique to deal with the problem above, because handling an image or photo that is damaged is also a process of manipulation of that image. One method of image repair due to noise is through a filtering process by dividing the image into four parts of the actual size, namely the filtered image and the other part are the details of the image where this process reflects the performance of the wavelet transformation method. The results of the trial using the software show the number of filter coefficients affecting the results of filtering with wavelet transformations, where the more the number of filter coefficients daubechies the smaller the visualization of contrast. Repeated decomposition of the image on the wavelet transformation will erode the original image, this is caused by the more pixel values transformed by both the low-pass-filter and high-pass-filter processes.