A software assistant for the design of realistic software phantoms

Abstract

Segmentation and quantification of medical image data are difficult problems in image analysis. Especially, an accurate and robust assessment of quantitative parameters is a key issue in many fields, and can have a direct impact on diagnosis and treatment planning. To this end, physical and software phantom data sets have become an integral tool during the design, implementation, and optimization of new algorithms. Unfortunately, a common research resource has not been established until now for many applications.In this work we propose a software assistant for the development of realistic software phantoms. Our aim is an easy-to-use tool with an intuitive user interface. Furthermore, we provide a software for researchers including a common basis of reference data, which facilitates a standardized and objective validation of performance and limitations of own developments, as well as the comparison of different methods. The fundamental phantom design focuses on objects that can be incorporated into a given background. This can either be a homogeneous artificial background, or a volunteer or patient data set. For each phantom the exact ground truth of the investigated object is available, which provides us with an excellent tool for the generation of realistic data sets. Several experiments are carried out for a number of different applications including software phantoms of small, hyperintense brain lesions, as well as software phantoms of liver metastases.