On evaluation of a multiscale-based CT image analysis and visualisation algorithm

Development of computed tomography (CT) protocols that minimise radiation dose for specific clinical treatments continues to be a major research focus. Building on the success of an earlier collaborative case study concerning prostate cancer diagnosis with pelvis CT images, this paper presents our evaluation results of a multi-scale texture analytic procedure developed to aid detection of specific anatomical features (abnormality, lesions, etc) in such images. This is a critical step in realising an intelligent and integrated image visualisation platform which will facilitate the construction of highly customised and personalised treatment plans. The ultimate aim is to provide in a single framework optimal software techniques for treatment planning, CT image-guided positioning and treatment delivery.