A new technique based on PIFS code for image retrieval system
I Ketut Gede Darma Putra, Adi Purnawan, Manuh Artana · Applied Mathematical Sciences · 2013
This paper proposes a new feature extraction method for image retrieval system based on local self-similarity PIFS code. The feature of image is represented by using distance and angle of a pair of range and domain block position from fractal code. This paper also uses variant contrast and spatial dimensions fractal feature for comparison. The testing of system’s performance is using 1000 images which have been stored in a database with 10 different categories with different number of image for each category. Images are classified as Homogeneous and Heterogeneous Image based on the complexity of the background image. The results show the proposed method gives a good performance.