摘要
提出结合àtrous算法与Retinex算法的医学图像增强.方法在二进小波理论基础上,利用推广的二维àtrous算法对医学图像进行处理,对经过多级分解后的医学图像的低频利用连续非线性函数增强,而对分解后的多个高频采用改进的Retinex算法进行增强;最后,对增强后的低频和高频进行重构.经过反复仿真试验结果表明:方法在去除噪声的同时更好地保留了图像边缘特征,增强了图像细节信息和凸显图像的轮廓,并克服对比度下降的不足,较好地去除光晕现象,视觉效果好,层次感更丰富.
Medical image enhancement is proposed with atrous algorithm and Retinex algorithm, the method in the dyadic wavelet theory based on the processing of medical images using a two-dimensional generalized atrous algorithm, on the low-frequency medical image decomposed by continuous multistage nonlinear function of multiple frequency enhancem-ent; decomposed by improved Retinex algorithm the final enhancement; to reconstruct the enhanced low frequency and high frequency. After repeated simulation test results show t-hat the method can remove the noise while preserving more image edge features, enhance image details, highlight the outline of the image, and overcome the shortage of low contrast, effectively remove halo phenomenon, visual effect well, layering richer.
出处
《数学的实践与认识》
北大核心
2017年第9期198-203,共6页
Mathematics in Practice and Theory
基金
国家自然科学基金(11261061
61362039
10661010)
新疆师范大学研究生科技创新基金资助项目(XSY201602010)
