仿生模式识别目标自适应跟踪算法

Research on biomimetic pattern recognition for object adaptive tracking algorithm

将仿生模式识别应用于图像目标的识别与跟踪,提出了一种能够自适应改变图像目标尺寸的目标识别与跟踪算法.仿生模式识别通过构建人工神经网络模型对训练样本进行覆盖,完成对不同尺寸目标的学习过程.针对仿生模式识别中有限数量训练样本不能够识别连续目标尺寸的问题,提出采用基于径向基神经网络的曲线拟合技术对仿生模式识别网络覆盖中的目标尺寸进行拟合,以实现目标尺寸的自适应改变;针对识别与跟踪过程中目标位置和尺寸的判定,提出采用基于欧氏距离的二次最小距离识别算法在图像序列中搜索目标.通过对比经典BP神经网络与文中算法,理论分析与实验结果均表明,该算法对于视频图像序列中目标中心和尺寸的识别是有效的和稳健的.

A new method for image object recognition and tracking based on Biomimetic Pattern Recognition （BPR）, which can automatically change the size of the object, is proposed. To accomplish the learning process of the sample objects with different sizes, an artificial neural network model is used to cover the training sample. The curve fitting method based on the Radial Basis Function （RBF） neural network is presented to approximate the size of the objects in the coverage of the network formed by BPR theory in order to change the size of the object automatically. The Quadratic Minimum Distance algorithm based on the Euclidean distance is applied to search the target in the process of object recognition and tracking. Experimental results and theoretical analysis show that the algorithm proposed in this paper is more effective and robust in object recognition and tracking of video image sequences than the BP neural network.