Digital image enhancement using iCAM

Abstract

Digital images taken with improper light exposure will contain no detail in shadow, or in highlight. The quality of these images could be enhanced using the image color appearance model, iCAM. One crucial step of the implementation of iCAM is in the step of image filtering. The aim of this thesis was to define the types of Gaussian filter used in the iCAM model. Two sizes of filter were varied: the size of the image size and half the size of image size. The sigma values of filter were varied between 1 - 500. This parameter varying was applied to the first filter used in the step of obtaining chromatic adaptation information and to the second filter used in the step of luminance adaptation. A combination of appropriate filters would provide good quality of images enhanced with iCAM. Three different types of digital images were tested. The results showed that filter size had little effect on the model’s performance. The image size was used as the filter size for both filters. The sigma of the first filter should be around 50, while that of the second filter is between 200 and 500. The use of iCAM with appropriate filters could improve quality of the input images in terms of image detail.