Abstract:
ABSTRACT Robotic vision is the main topic of research nowadays, in order to reduce the computation time while making visual decisions, it’s very important to visualize a scene according to human vision. Human visual system focus on certain regions/objects of particular interest while ignoring the surroundings, our project extracts those regions of interest according to visual senses/behavior of human being. The purpose of this project is to design and develop a software system that takes dynamic and static scenes as an input and extract the salient regions based on human visual attention system. This project presents a model of visual attention that can be used as a building block for a real-time robotic vision system. Visual attention helps us decide what's important, and understanding how this works in humans may allow us to build smarter computer vision systems that know what to look for. Region based approach is used to reduce the number of units that have to be processed by the attention systems. Top-down, bottom-up techniques and faster algorithms that are used that will help in increasing performance and efficiency of processing to integrate the attention procedure into a comprehensive biologically inspired vision system. Five features namely color; size, orientation, eccentricity and symmetry are used.
