Nonlinear ill-posed problem in low-dose dental cone-beam computed tomography

Abstract This paper describes the mathematical structure of ill-posed nonlinear inverse problem low-dose dental cone-beam computed tomography (CBCT) and explains advantages a deep learning-based approach to reconstruction images over conventional regularization methods. underlying reasons why CBCT is more than standard tomography. Despite this severe ill-posedness, demand for systems rapidly growing because their cost competitiveness low radiation dose. We then describe limitations existing methods in accurate restoration morphological structures teeth using data severely damaged by metal implants. further discuss usefulness panoramic generated from tooth segmentation. also possibility utilizing radiation-free intra-oral scan as prior information image compensate damage caused