This repository contains the presentation and report submitted for the seminar on the paper: A Model of Saliency Based Visual Attention for Rapid Scene Analysis by Laurent Itti , Christof Koch and Ernst Niebur, IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 20, NO. 11, NOVEMBER 1998
Summary: This paper discusses the saliency based model of attention which is based on Human visual strategy. In case of the human visual system, it is believed that the feature extraction happens at an early stage and the semantic information is gathered at a later stage when we are presented with an image. Hence, the visual system focuses attention to those locations constituted by separable features. The model discussed here is developed to mimic this behavior. The image is decomposed into a set of feature maps. The locations that stand out from their surroundings in each feature map are fed to a "saliency map" which guides the shift in attention. A neural network outputs the locations in the decreasing order of saliency.