针对现有小波轮廓描述符依赖于起始点的问题，提出一种自相关小波轮廓描述符。首先对经过平移和尺度归一化后的轮廓进行小波变换，再将各个频段小波系数进行自相关变换，消除起始点变化对描述结果的影响，得到最终的自相关小波轮廓描述符。该描述符彻底摆脱了对起始点的依赖，同时具有较高的描述精度。实验结果证明了自相关小波轮廓描述符的有效性。

A deficit of common wavelet contour descriptors is the dependence on starting point, and an autocorrelation wavelet contour descriptor is proposed to solve the problem. In the scheme, the wavelet transform of contour is carried out after translational and scaling normalization, then autocorrelation transform of wavelet coefficients at each bandwidth is implemented, which eliminates the influence of starting points varied in describing, and then an autocorrelation contour descriptor is achieved at last. The proposed descriptor which realizes the independence of starting point could also maintain high accuracy. Experimental results confirm the efficiency of the descriptor.