基于多尺度二维直方图均衡化的医学图像增强方法

Multi-resolution based two-dimensional histogram equalization for medical image enhancement

二维直方图均衡化(2DHE)算法可以使图像对比度在一定程度上得到提升,但是容易丢失图像局部细节,尤其是在高比特的医学图像中更易丢失.针对此问题,提出一种全新的医学图像增强算法模型(Lap-2DHE).该算法采用Laplacian金字塔对图像进行细节增强,然后用二维直方图均衡化算法对图像进一步增强.仿真实验结果表明:新方法不仅解决了原有2DHE方法中存在的细节丢失问题,还为包括高比特成像方式在内的医学图像提供了一种普遍可行和有效的处理方法.

In this paper,we shall present a new medical image enhancement method by developing two-dimensional histogram equalization algorithm(2DHE).As known,2DHE can improve image contrast effectively,but also damages imaging details when applying it to high bit-depth medical images.For solving this problem,we integrate Laplacian pyramid with 2DHE algorithm for detail maintenance and put forward a multi-resolution based two-dimensional histogram equalization algorithm(Lap-2DHE).The experimental results show that the new method not only overcomes shortcomings of the existing 2DHE in the disposal of the details,but also provides a generally feasible and efficient way for medical images,including high bit-depth imaging.