计算机科学与探索 ›› 2011, Vol. 5 ›› Issue (11): 1014-1020.

• 学术研究 • 上一篇    下一篇

改进视皮层视觉机制的视觉注意力模型

韩 冰, 高新波, 李 洁   

  1. 1. 西安电子科技大学 电子工程学院 影像系统实验室, 西安 710071
    2. 西安电子科技大学 电子工程学院 智能感知与图像理解教育部重点实验室, 西安 710071
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-11-01 发布日期:2011-11-01

Visual Attention Model Based on Visual Cortex Mechanisms

HAN Bing, GAO Xinbo, LI Jie

  

  1. 1. Video & Image Processing System Lab, School of Electronic Engineering, Xidian University, Xi’an 710071, China
    2. Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education, School of Elec-tronic Engineering, Xidian University, Xi’an 710071, China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-11-01 Published:2011-11-01

摘要: 人类的视觉注意机制是人类大脑感知事物的最直接的功能。提出了一种基于视皮层视觉机制的生物激励注意模型。利用HMAX(hierarchical maximization)模型的四层机制中的C1细胞单元图, 构造独立成分分析(independent component analysis, ICA)滤波器组, 进一步利用对尺度、平移等均具有不变性的C2细胞特征, 以及香农熵理论, 共同构造用于视觉显著性区域检测的测度。在心理学实验的自然场景图像以及心理学刺激模式上的结果表明, 该方法与传统方法相比, 更符合人眼的感知特性, 从而进一步验证了该方法的有效性和准确性。

关键词: 视觉注意机制, 显著性图, 视皮层, HMAX模型, 人类视觉信息处理系统

Abstract: The visual attention mechanism of human brain is the most direct function for human perception. This paper proposes the biological inspired model based on visual cortex. Using the C1 cells in four layers structure of hierarchical maximization (HMAX) model, it constructs the independent component analysis (ICA) filters. Fur-thermore, it also presents the measure of saliency detection by combining the C2 cells, which have the good charac-teristics in scale and translation invariance, with Shannon theory. The experimental results demonstrate that the pro-posed method has good performance over classical algorithm on real nature scene dataset and psychology stimuli patterns.

Key words: visual attention mechanisms, saliency map, visual cortex, hierarchical maximization (HMAX) model, human visual information system