基于视觉注意机制的图像分类方法

Image classification based on human visual attention mechanism

针对Itti视觉模型在模拟人类视觉的过程中没有考虑到人眼视觉系统对外界事物边缘和轮廓的注意敏感性的问题,提出一种改进的基于视觉注意机制的图像分类方法,即基于Itti模型,同时引入图像的边缘特征对图像进行分类.经实验验证,与传统的Itti模型相比,该方法具有更高的分类准确率,分类准确率达到97%以上.

To solve the problem that Itti visual model does not considering the sensitivity of the human visual system on the edge and contour of the external things when simulating human vision, an improved image classification method based on human visual attention system is put forward. On the basis of Itti visual model, edge features are introduced into image classification. Experimental results show that compared with the tradition model, this novel algorithm has a higher classification accuracy over 97%.