彩色图像误差扩散多值量化滤波器的优化设计

Optimized Design of Multilevel Quantization Filter in Color image Error Diffusion Alogrithm

在多值量化情况下 ,因量化误差为白噪声的理想条件无法满足 ,故实施误差扩散处理时仅单纯地进行高通滤波 (误差蓝化 )将损害图像固有的低频特性。误差扩散算法中较为理想的误差滤波器应在高频域具有较大增益 ,而在低频域具有有限度衰减。在兼顾滤波器稳定性因素的同时对误差滤波器的脉冲响应进行了重新设计 ,并以此构成了用于彩色图像颜色量化的多色调误差扩散优化处理算法。

Because the ideal condition which quantization error has a white power spectrum is never valid on the state of multilevel quantization, the intrinsic low frequency characteristic of image is inflected by the processing of error diffusion only dealed with highpass filter (error blue shift). In this paper, it is proposed that the ideal error filter of error diffusion alogrithm should have greater gains on the high frequency, and finite attenuation on the low frequency. Accoding to such idea, the impulse response of error filter is designed anew, and the stability of filter is taken into account. With the filter the optimized algorithm of multilevel error diffusion used for color image quantization is formed.