| dc.description.abstract |
In addition to the color representation contained in a digital image, the size of the image will also greatly affect the image processing process. In some cases of image acquisition, sometimes the size of the image obtained does not correspond to the desired one. Pattern recognition processes, such as vehicle license plate recognition, may result in incorrect output if the size of the processed input image is not appropriate. Therefore, determining the size of a digital image is very important in further image processing. To solve such problems can be done by enlarging the size of the image through the process of interpolation or resampling. Image magnification is the process of enlarging the size of a digital image from its original size to a new and different size based on the desired magnification scale. However, as a result of enlarging the image obtained there is always a blur, so the image will look like a checkerboard (split). This situation occurs because during the magnification process the resolution of the new image becomes lower. Thus, additional methods are needed to improve the image quality of the resampling results. This study aims to improve the low-resolution image quality of resampling results by applying Linear Interpolation and Geometric Mean Filter methods. The test results showed that the Linear Interpolation method was able to produce a new image that was larger than the original image and the Geometric Mean Filter method could improve the image quality of the enlarged image so that the details of the information in the image are better and clearly seen visually. And to find out if the image enlarged from the Linear Interpolation and Geometric Mean Filter is good or not, we will use the Mean Squared Error (MSE) and Peak Signal to Noise Ratio (PSNR) methods.
Keyword : Image, Linear Interpolation, Geometric Mean Filter, MSE, PSNR
| en_US |
