A new filter bank algorithm for enhancement of early signs of ischemic stroke in brain CT images

Brain CT images are very useful in diagnosis of cerebrovascular accidents (CVA). These images contain too many information. But most of the times, provided information is contaminated by noise and suffer from poor contrast. On the other hand, there are certain parts of the brain image that is really important to radiologist, while other image details are more or less confusing. This sparks the need for customized enhancement of brain CT images. Translation-invariant wavelet transform is being widely used in most of image processing tasks including image enhancement. This transform is calculated with a filter bank algorithm, called algorithme àtrous. In this paper we propose a filter bank structure similar to algorithme àtrous. This structure is more redundant and offers greater selectivity and flexibility to enhance desired features of brain CT images.