Volume Cutting of Medical Data Using Deformable Surfaces Modeled with Level Sets

With the advancement of digital imaging technology in the medical domain an increased amount of sampled biological data is being generated. Clipping of volume data has become more and more important because it allows us to cut away selected parts of the volume and plays a crucial part in medical image understanding, computer assisted diagnosis and surgery simulations. We propose a method for volume cutting using deformable surfaces modeled with level sets. Using Boolean operations the method is extended to multi-object clipping. The overall computational cost is reduced by using a fast and computationally efficient narrow band level set algorithm. The proposed model has been used to extract arbitrary shapes from scanned volume data including some low contrast medical data with promising results.