一种基于改进视觉注意模型和局部自相似性的目标自动检测算法研究

针对基于视觉注意模型的检测算法只能检测到图像中的感兴趣区域,无法准确地提供目标的轮廓和位置的不足,提出了一种基于改进视觉注意模型和图像局部自相似性的目标自动检测算法。通过增加运动速度和运动方向特征改进了经典的Itti视觉注意模型。利用改进的视觉注意模型提取感兴趣区域,提高了视觉注意模型的检测能力。再利用图像在边缘处具有良好的局部自相似性,实现了基于图像局部自相似性的目标检测算法。实验表明,算法能快速检测到图像中的目标感兴趣区域,并对其进行精确分割和定位。

Research on target detection algorithm based on improved Visual Attention Model and local self-similarity character

Aiming at the problem that the target detection algorithm based on visual attention model can only detect the Regions Of Interest (ROI) in image and it can not get the outline and accurate position of target, this text propose a kind of target detection algorithm based on improved visual attention model and local self-similarity character. It has improved Itti vision attention model through adding movement velocity and movement direction character. The detection ability has greatly advanced by using improved visual attention model. After getting the interested region, it realizes the target segmentation by using the good local self-similarity character in the edge of target. The experiment result shows that the new algorithm can quickly detect interested region of target in picture and position it accurately.