Abstract: The digital medical image has its own particularity，in order to diagnose correctly disease，the medical image need be enhanced.According to the particularity，a similar method of Retinex to enhance the images is proposed.In the method，the images are smoothed by a smoothing conduction function which consists of gradient and context.The images are enhanced by the method，and the enhanced images are operated twice by transformed space stretching.Experiments and analysis show that the method is obviously better than Histogram Equalization，Homomorphic Filtering and multi-scale Retinex.

Key words: digital medical image, Histogram Equalization, Homomorphic Filtering, multi-scale Retinex, smoothing conduction function, similar method of Retinex, transformed space stretching

摘要： 数字医学图像有其特殊性，为便于医生的正确诊断，需要对图像进行增强处理。根据数字医学图像的特点，提出了类似Retinex方法。该方法采用梯度和context构成的平滑函数平滑图像，利用类似Retinex原理增强图像，并对增强后的图像施行二次变空间拉伸。实验及主客观分析表明：该方法对数字医学图像增强的综合效果明显优于直方图均衡、同态滤波和多尺度Retinex方法。

关键词: 数字医学图像, 直方图均衡, 同态滤波, 多尺度Retinex方法, 平滑传导函数, 类似Retinex方法, 变空间拉伸, ,