A soft-computing-based approach to artificial visual attention using human eye-fixation paradigm: toward a human-like skill in robot vision