Regionally Adaptive Histogram Equalization of the Chest

Advances in the area of digital chest radiography have resulted in the acquisition of high-quality images of the human chest. With these advances, there arises a genuine need for image processing algorithms specific to the chest, in order to fully exploit this digital technology. We have implemented the well-known technique of histogram equalization, noting the problems encountered when it is adapted to chest images. These problems have been successfully solved with our regionally adaptive histogram equalization method. With this technique histograms are calculated locally and then modified according to both the mean pixel value of that region as well as certain characteristics of the cumulative distribution function. This process, which has allowed certain regions of the chest radiograph to be enhanced differentially, may also have broader implications for other image processing tasks.