Selection of Visual Attentive Area Based on Scale Space Representation
Jing Shao · Microelectronics & Computer · 2009
Contraposing the multiple-scale representation of visual feature image in the computational model,a visual feature image representation method based on non-linear scale space is proposed.In this case,not only the center-surround computational policy is realized,the specific information,such as edge,can be reserved efficiently.Besides,contraposing the optimal size selection of the salient region,a computational method of optimal size selection of salient point based on scale space representation is researched.The validity and rationality in cognition of the algorithm are demonstrated through the visual search experiments.