4D Lung Reconstruction with Phase Optimization

This paper investigates and demonstrates a 4D lung CT reconstruction/registration method which results in a complete volumetric model of the lung that deforms according to a respiratory motion field. The motion field is estimated iteratively between all available slice samples and a reference volume which is updated on the fly. The method is two part and the second part of the method aims to correct wrong phase information by employing another iterative optimizer. This two part iterative optimization allows for complete reconstruction at any phase and it will be demonstrated that it is better than using an optimization which does not correct for phase errors. Knowing how the lung and any tumors located within the lung deforms is relevant in planning the treatment of lung cancer.