Seyyed Mohammad Reza Hashemi - Health Technology Incubation Center,Semnan University of Medical Sciences Semnan, Iran
Mostafa Mohammadpour - Health Technology Incubation Center,Semnan University of Medical Sciences Semnan, Iran
Ehsan Kozegar - School of Computer Engineering Iran University of Science and Technology, Tehran, Iran

Photographers used a bunch of photography com positionrules to create high-quality and more attractive photos.In this work some rules are used to determine which place inan image attracts human attention. We present a novel saliencydetection method, which utilizes the knowledge of photographiccomposition to improve saliency detection results. For this purpose,MIT eye tracking dataset was utilized in which observer’seye path through the image was obtained and recorded in aground truth file. Then, low-level features such as orientation,color, intensity are combined and train a SVM classifier froma direct mapping those features to eye fixations. Comparisonbetween proposed model with some of state-of-the-art models wedemonstrates that our model reveal salient region in image whichis important for photographer’s view.

Photography Composition Rule, Attention, SaliencyDetection, Eye Tracking, SVM