Landscape Image Filtering via Aesthetic Inference
Mahsa Ghane · IOSR Journal of Engineering · 2012
Today's because of increasing photography facilities, produce a high number of images.Hence, we need devices and applications for better images search and stronger management of them based on qualities or quantity of aesthetic.These help us to create digital image albums and auto management of contents according to aesthetic concept.In this paper first, we want to introduce the general features in aesthetics in different scopes of photography such as portrait, macros, news, sports, animals and etc.Second, trying to determine the particular influence of each feature on different scopes of photography because aesthetic features are presented in photography are very general.Next, among of these provided features, the top six features that specifically have the highest influence on the landscape images are proposed.Accuracy verification of these features was shown by comparing images from a standard dataset.Supposing the ranking of aesthetic in the dataset from 1 to 7, 1(very ugly)and 7(very beautiful), it is observed that proposed features was influence due to their ranking in dataset were more than 5 and none of them had ranking scores with less than 5 in the standard dataset..